CICS User Group | October 2026

Using CICS Security Discovery to adopt a zero-trust stategy

Many CICS environments continue to rely heavily on transaction-level security, often resulting in users retaining access privileges that are no longer appropriate for their current roles. This session will provide an end-to-end overview of the CICS Security Discovery workflow and demonstrate how the tooling can help organizations better understand and modernize security configurations within production CICS regions. Attendees will learn how Security Discovery analyzes existing security definitions and production usage patterns to identify user roles, application access, command usage, and resource permissions. The session will also include a live demonstration of the Security Discovery Explorer Editor and show how organizations can generate actionable to-do lists to support security modernization and zero trust initiatives.

Read the Transcription

[00:00:09.] – Amanda Hendley
Welcome. If we haven’t met, my name is Amanda Hendley. I am your host for today’s Mainframe Virtual User Group, and today we’re here talking CICS. So thank you for joining us. This is actually one of 4 user group sessions this month. If you attended earlier in the month, we are going to have a redo of that session. It is just about to be scheduled, and we’ll see you there in a few weeks. But for now, let’s, let’s get started. Again, I’m with Planet Mainframe, and I want to thank our partners for sponsoring this series. Broadcom Mainframe Software and Data Kinetics are our CICS sponsors, and we’re here today for CICS Security Discovery with John Taylor, who is joining us from— well, he’s with IBM Hursley, and I think he’s joining us from home today. So glad to have you here, John. Just so we can get to know you and learn a little bit about you, if you’ll drop your location and job title in chat, that would be great. It’ll also let me know that our chat functionality is working. So later when you have questions, you will know how to ask them via chat.

[00:01:34.] – Amanda Hendley
And that’s the next thing I’ll mention is that we’re going to do questions today as we go. So you can drop yours in chat at any time. And additionally, my email address is ahendley@planetmainframe.com. If you have anything you want to share with the CICS community or you want to put out there a request for content. So I do hope you’ll drop your job title and location in chat. I’m gonna do mine right now. I’m a managing editor, and I am— I’m in a suburb of Atlanta, Georgia, Marietta, if you’re familiar with it. The other day I was actually on one of these user group sessions, John, and someone was like 20 minutes from my house. Like What are the chances? It’s so awesome. So thank you again for being here. We’re gonna just do a quick shout out. We’re here for a security session, so you might be interested in the security content we’re publishing all month long at Planet Mainframe. And as I was telling John earlier, we did it in October to kind of like do the whole like, hey, it’s October, Halloween, security scariness. And I’m having scary IT issues this past couple days.

[00:02:52.] – Amanda Hendley
So yay for Security Month. We’ve got some great content on the website. We’ve been publishing, gosh, now we’ve probably put about 12 new pieces of security-focused content up there at planetmainframe.com. So please check it out. And lastly, I will make a mention that the Arcati Mainframe Navigator, which is our research report, we’re getting ready to release the survey. So you’re gonna start seeing emails from us about that survey, and we’d love to hear from you. It’s real important that we get all the opinions and viewpoints and data in so that we can write a report that’s valuable to the entire community. So with that, enough talking from me. I’m gonna invite John Taylor to share screen, and I’m gonna stop my share, and his should be sharing already. Let’s see.

[00:03:47.] – John Taylor
Is it sharing?

[00:03:49.] – Amanda Hendley
Is it sharing? It is sharing. So let me introduce you. John is an IBM software engineer and CICS developer working at IBM Hursley. As I mentioned, he specializes in modernization and system performance and is gonna talk to us today about CICS. Security Discovery. I’m going to turn it over to you.

[00:04:09.] – John Taylor
Cool. Thank you for that one. Right. So before we start, a couple of things to mention. If you’ve got any questions, please fire them out straight away. It’ll stop me talking so much, which would be a bonus for everybody. If they’re really, really difficult questions, thankfully we’ve got Mark Bocock on the call as well, and he can do the really hard stuff. If there’s anything easy, I’ll handle it. Mark can do the rest. So let’s get past the first slide. Probably worth mentioning that I will be sharing the slides after the session, but a big chunk of the session will be a demo, which obviously won’t be in the slides. So hopefully it’ll be interesting and we’ll see what we get. So titles all well and good. If I can just make the page down button work, I’ll be much happier. There we go. Right, I don’t think I really need to focus too much on the fact that there’s cyber issues out there that we need to worry about. I’ve seen one or two mentioned in the news recently, so I think suffice to say that there are lots and lots of cyber issues out there, lots of things we have to worry about.

[00:05:01.] – John Taylor
Now I think the one thing that’s worth mentioning particularly today because we’ve got all this, all these concerns about PHP, is while certainly a lot of hacks and exposures happen because of problems such as, well, as it turns out, PHP being a bit flawed, there’s also a lot of problems that happen just because of simple use of valid user credentials. So escalation of authorities, things not being controlled as well as they should be. And a lot of what I’m going to talk about today is trying to address that style of problem. And we’ve got some stats there, some numbers which are possibly on the low side. But I think suffice to say that a breach can be very expensive, both to your business and to your reputation. So Zero Trust, great philosophy. The key thing about Zero Trust is we’re trying— we’re thinking about worrying about protecting the resources within your system, not just protecting the boundaries. There’s this almost catchy phrase here, enabling the right users to have the right access to the right data for the right reasons. Quite lengthy. But very simply saying, we really want to be sure that you’re accessing something, only things you should be allowed to access, and you should be accessing it for a good reason, for a reason that’s correct.

[00:06:15.] – John Taylor
Another key point is don’t trust things, always verify. So just because somebody’s got inside your boundary or they’ve logged into VPN, you don’t assume they should access everything. Don’t assume that people are who they claim they are, etc. And a good philosophy is to always assume that the bad actor’s already present and always be continuously monitoring just so that when things do happen, you’re aware of them sooner rather than later. Now, got this lovely jigsaw diagram here which shows 6 different areas. Today, I’m really only focusing on 3 of them. So identities, who is this user, who is this task ID, the applications that access the data for those identities, and the data itself. So we’re less worried about infrastructure, network, and devices today. They’re all very important. They’re just not what I’m talking about in this particular presentation. So I’m assuming that everybody on the call is at least fairly familiar what CICS is and has some experience of how we tend to configure CICS. And historically, it used to be the case that you would enable transaction security, at least on your production regions, and you say, okay, you can only do stuff if you’re authorized to.

[00:07:28.] – John Taylor
to drive the relevant transactions. And as long as we make sure that you can only drive the transactions you’re allowed to be— that you should be allowed to drive, everything will be good. And that used to be probably okay, but nowadays that’s not necessarily going to be good enough. And transaction security is arguably very much a matter of boundary security, but we’re certainly nowadays worrying a lot more about security in depth. So not just are you allowed to run the transaction, but are you allowed to touch the file that that transaction touches? Are you allowed to use the TSQ, TDQ, whatever? So there’s an opportunity there for securing much more than just which transactions is a user allowed to run. But of course, figuring out exactly what security settings you need to switch on can get very, very difficult, particularly in a system that’s been running a long time, that’s built up over time with lots and lots of users. And I’m hoping that at the end of this presentation, you’ll say, oh, there’s lots of ways I can address those problems. It’s not an insurmountable challenge to switch on file security, say, in my CICS regions.

[00:08:34.] – John Taylor
And that’s very much what I’m hoping to show you is how you can improve your existing security definitions and how you can create new security definitions to make you more secure overall. And key to all that is roles. So It’s a shimmy. Certainly on some systems that I’ve talked to customers about, some systems I’ve used, there’s been a problem that somebody will join, will join the department and they’ll say, oh, I need access to do this and I need to do this new job. And I said, well, what access do you need? They say, well, I’m doing the same job as Bob sat next to me. Just give me the same access Bob’s got and everything will be great. And that does work. It does work. But then you move to another role and you gain some more permissions and over time you gain lots and lots. And when you leave a role, it can be very difficult to say which permission should I take away from this person? And this just builds up over time and it can become a real challenge. So the phrase that we had earlier, enable the right user to have the right access to the right data for the right reasons.

[00:09:35.] – John Taylor
Well, that’s pretty hard to do if you’re just granting individual access to individual people. A much better approach is to define a role and say, okay, this role is, say, I’ll take a simple one, an operator. The operators need access to certain transactions, to certain resources. And we define all those permissions carefully for operators. And then we say, okay, if you’re an operator, I’ll put you in that role. And the nice thing is, if that person leaves the role of operator, maybe to go off and be something like Sysprog, or maybe become something less useful like a manager, then obviously you want to take away the permissions they had as an operator. Now, previously that was quite hard because you have to unpick the permissions, say, which should I take away? But if you’ve just put this person into a role and you’ve granted permissions to the role, it’s very simple. You take them out of that role and you put them in whatever role they need for their new job. So that’s why roles are great. Now, problem is that if you haven’t already set up roles from day one and got everybody in the right role, again, it can be really, really hard to figure out what the role should be, which users should be in it, which permissions should be in it.

[00:10:44.] – John Taylor
And again, that’s something I’m hoping I can show you shortly. Oh, it’s also probably worth mentioning the dreaded auditors. It can be very difficult to prove to an auditor that you’ve got exactly the right permissions for all your users if you’re granting permissions to individuals. So again, that’s worth mentioning. Let’s see what my next slide is. I’m really trying to get to the first presentation because it’s much more fun doing presentations than slides, but I want to give a little background. So the Security Discovery Editor is the main thing I’m going to be showing you today. So I will get around to it in just a couple of slides, I promise. Couple things to mention about it. The Security Discovery Editor is designed to work with your production systems. This is not something that we say, oh, use it on your test regions, but don’t dare use it in production. The intention is that you work with your production systems with this one. Now, we do make a couple of presumptions. We assume that you have security active. If you don’t have security active, then, well, you’ve probably got bigger problems than I can solve with the editor.

[00:11:43.] – John Taylor
And as a Basic starting point, we assume you’ve got transaction security already enabled, that you’ve locked down your transactions at least moderately well. And that’s because that gives us a useful starting point. Now, if you haven’t got those, then we can certainly have a chat about how to get you to that position. But that’s not really what the security is aimed at. We’re aimed at this assumption that you have got reasonable security, which is perhaps difficult to manage, which you want to extend, which you want to make easier to manage. And that’s what I’m hoping I can help you with. And the other thing I really want to do is help you to extend the security definitions you’ve got today to cover more resources within your CICS TS to keep the system more secure overall without giving you an enormous increase in workload just managing everything. So can’t say I’m crazy about this slide, but it is quite useful, this picture. And I’ll take a couple of minutes because I’m going to use this picture a fair bit in some of the subsequent slides. So it’s kind of handy. And I’ve got this proportionately large box here called CICS Explorer, which is— the CICS Explorer contains the Security Discovery Editor.

[00:12:51.] – John Taylor
Now, the nice thing about that is that I believe that most CICS Sysprogs have CICS Explorer installed. If you don’t have CICS Explorer installed on your local workstation, well, it might be a really good thing to get because it’s a really powerful tool. It does lots and lots of things. And The CICS Security Discovery Editor is one of the perspectives in CICS Explorer. The other nice thing about it being an explorer is as we keep updating the editor and adding new features to it, you merely need to grab the latest version of CICS Explorer and there you go, all the features are there. So there’s several things that the editor uses input and something it uses output. Specifically, the starting point with the editor is we use data from your RACF. Now, we, with CICS, we ship some simple JCL that runs in REX. We did that specifically in REX so you can have a good look at what we’re doing, see we’re not doing anything fishy. And all we’re basically doing is having a look at your existing RACF definitions, and we’re converting them into a YAML file, so a semi-human-readable file that can then be read into the CICS Explorer easily to have a look at exactly what you’ve currently got defined.

[00:14:03.] – John Taylor
And there’s quite a lot you can do with just that data. And the first demo that I’ll show you is purely using that data from RACF. Now, the nice thing about that is that whatever version of CICS you’re using, that’s going to work for you today. Even if you’re running on a version of CICS, say, I don’t know, CICS 5.6, for example, you could grab an install of CICS version 6.x, grab the REXEXECs from there, grab the JCL from there, and do an extract from RACF to start looking at your data, start improving your data, even if you’ve not yet moved on to one of the more recent versions of CICS. Now, the second demo in the second part of the presentation, I’ll just mention it briefly now, we’ll use some data extracted direct from your CICS regions, and that’s going to require CICS version 6.2 or later, but we’ll worry about that one in a bit. When you’ve edited your data, in the CICS Explorer, in the Security Engineering Explorer, at some point you’ll say, I think these definitions are way better than definitions I used to have. So at that point, there’s a couple of things you can do.

[00:15:04.] – John Taylor
You can compare the definitions that you’ve now got with the previous version of definitions, just have a quick check, see if anybody’s gained anything you wouldn’t really want them to, or if somebody’s lost a really critical permission in your shuffling of things around. And you can also export the data from the Security Editor as again a YAML file, or as we call it, a .ESM file, just to identify the file type uniquely. And that ESM file, again, we provide a REX job which will convert the YAML into REX commands— sorry, into RACF commands to create definitions in your RACF that you can then start using in preference to your old definitions. Now, depending on how risk-averse you are, or just how much you want take huge risks. You’d probably want to switch on warning mode just in case you’ve made a mistake with those definitions initially when you bring the CICS TS up, try using them, make sure everything’s happy. If you don’t switch on warning mode and you’re really confident you’ve done it right, well, either it’ll all be brilliant and everybody happy, or your phone will start ringing a lot and people will start screaming about the fact they can’t do their day job anymore.

[00:16:10.] – John Taylor
So warning mode can be nice just to make sure that you’ve not changed anything you shouldn’t. So that’s in very, very broad terms though, the overall process. But I’d like to go into a little bit more detail about a couple of bits. I mentioned that when we extract the data from, from RACF, we have a simple bit JCL, which you can see, you can see how incredibly simple it is at the top right here. And that runs a series of commands against RACF and generates a YAML file which you can see an example of here, nicely color-coded to highlight it as well. Now, extracting that data from RACF will require RO audit authority against RACF, so you may have to ask your systems— sorry, your security administrator to run the REX for you just to have the authority to do it. But again, the fact that it’s JCL running REX, it means everybody can have a good look at it and see there’s nothing fishy going on here. This is just basic RACF commands extracting data and simple Rex formatting it into an approximately YAML format, just so it’s something you can easily read into the CICS Explorer.

[00:17:22.] – John Taylor
And once you’ve got that data into CICS Explorer, at that point you can start doing all the cool stuff. You can start looking at your data, you can say, okay, what needs to be improved here, what looks great for that matter. You can start to get a good idea of what your data looks like. And I have got a picture of the editor shown there. Be much more fun to show the real editor live in a moment rather than static pictures. Static pictures are boring. We’re nearly at the first demo, which I’m pleased to say. The one other thing I wanted to really highlight here was, as I said, anytime along the process that you’re editing this file, you might be thinking, well, okay, I know I’ve made a bunch of changes. Have I done anything really stupid? Have I taken away some really important authority everybody’s going to need? Have I accidentally granted way too much authority to people they didn’t previously have? And a little while ago, we added a feature to the editor to allow you to compare whatever you’ve got in the editor right now with any ESM file that you have.

[00:18:24.] – John Taylor
Now, typically that would be the ESM file you imported in the first place, and if that’s the case, you don’t even have to specify the file because I’ve kept a copy of it. Or you can just say, well, I’ve got this ESM file that I generated at some point, I want to see where we got to. It’s a nice way of just having a quick sanity check of have you granted any authority you shouldn’t or taken away any you shouldn’t. So without further ado, I have my first favorite slide, the one that says Demo 1. So, so far the key point here is we— this is just using data from the ESM file from RACF. So doesn’t matter what version of CICS you’re running, you can do this part of the demo and a few other things besides. right now with your CICS regions if you so wish. So I’ll move the slide out the way. I hope everyone can see my screen clearly. So what we have here is the CICS Security Discovery Editor. As I said, it runs as part of CICS Explorer, so it’s an Eclipse-based application. So there’s certain things you’re probably familiar with about it, but I’d like to draw your attention to certain broad aspects of it before I go in too deep.

[00:19:31.] – John Taylor
In the center, I’ve got a table here. And which one? Oh yes, to-do list. Excellent. So I’ve got a table that shows these. I probably should mention everything I’m going to show you today is made-up data from our test systems. There’s no sensitive data here. There’s nothing here that I’m going to get in trouble with GDPR for. It’s just test data that we’ve used in various tests. Now that has the bonus that some of this data is quite entertaining and quite silly. But I’ll draw your attention to some of those in a moment. The center table, it shows you I’ve got lots and lots of users down the— as the rows down the left-hand side. And you’ll see that some of these users are in imaginatively named roles such as Group 3, Group 4, etc. And I’ve also got several users under this area called unresolved. So that will be users that are not in a particular role yet. So we might want to improve that in a bit. And also along the top, I’ve got some incredibly questionably named member lists and generic profiles, and even a few resources, in this case transactions. Now, I think it’s probably worth mentioning at this point, this particular piece of test data was created by one of my colleagues when we introduced the to-do list, which is on the left.

[00:20:52.] – John Taylor
And the to-do list highlights things that might be bit wrong in your system, or possibly a lot wrong in your system. And the things that we think you might want to have a look at and say, should I improve this? And my colleague, he’s a superb tester. He tried to make every single concern that we could raise in the to-do list appear in a single file of data. And he got pretty close until we kept adding new to-do list items. So I hope that nobody will ever see a to-do list this long. against real production data. And I hope that nobody will have any data look quite as bad as this. But the idea is it still holds in that we’ve got data here that says these are the users accessing these resources. And you’ll notice we’ve got some Rs and some As here. Quite imaginatively, R means read, A would be for alter, and they could also be update and control. Now, at the moment, we’re looking at transactions. So One thing you’ll point out to me is, why have you granted alter authority to a transaction? And the simple answer is that was to get another thing to flag up in the to-do list as looking wrong.

[00:21:59.] – John Taylor
I’d like to mention that down the right-hand side, whenever you click on a cell or a user or a resource, we will show you details about the thing you’ve clicked on. We will tell you what we know about it. So if you need to do a deep dive, try to figure out exactly what permissions a given user’s got, or maybe which roles a given user is in, for example. You can find out that sort of information in the details list. A lot of the time you don’t need that level of detail. When you do need that level of detail, it can be really, really handy to see it there. Now finally, probably worth mentioning, there’s a really boring looking gray area at the bottom here which you won’t see until we hit It won’t get any more interesting until the second demo because that requires the data that we extract from running CICS regions. So enough about the general layout. I want to go back to the to-do list. Now, as I said, this is hopefully the worst to-do list anyone will ever see. And this shows you a lot of the things that we call out.

[00:23:02.] – John Taylor
So we scan your data. So this is personalized to your particular setup. And it says, these are things you might want to worry about. And some of these are really, really obvious. Like invalid role names, quite a nice example of something that’s really obvious. And I’m sure you know in RACF, a role name will be made up of uppercase characters only. But because this test data, we managed to pull it in with some lowercase characters. And there are ways of getting that, which I’ll show you later, quite legitimately through using the security discovery. But it’s worth pointing out that you’ve got this invalid role name. if you tried to import the station into RACF, you’d have a problem because it’s lowercase. So you’re telling me that there’s a role name that’s invalid. And if I click on it, it will give you a bit of details and say, well, what’s wrong with it? It said, well, yeah, it’s not valid. Got to change to a valid name. And we offer some actions, which unsurprisingly with an invalid role name, there’s not a whole lot you can do. I can highlight it for you if it wasn’t already highlighted, and I’ll give you the option of renaming it something sensible.

[00:24:06.] – John Taylor
And I’m not sure it’s necessary since it’s a role name, but there you go. We’ve now got one less invalid role name. Fairly trivial one. More interesting ones that we’ve got are things like— here we go. I’ve got a member list with universal access other than none. So fairly reasonable rules I think most people agree with is that UAC other than none is not something you would tend to want to use. There are very few CICS resources or even CICS transactions that you would say absolutely every user on the system should be allowed to run this transaction. I mean, perhaps help, but beyond that, there aren’t too many transactions that you’d want all users to be able to run. In fact, they’re bordering on none. So one of the things we will call out is when we find use of UAC other than none and say, is that the best thing what to do? Should you really have every single user on this system, irrespective of who they are, being allowed to run this transaction, whatever it is. And again, we highlight that briefly in the description and we have a couple of options for you.

[00:25:13.] – John Taylor
I can just make it— I’ve got a little cursor hiding in the way. I can get it to show me the member list that has the problem, or I can say, no, I want to convert this universal access because you’re quite right, I shouldn’t be granting this authority to absolutely everybody. There’s actually a much smaller subset of people that should have authority to run this these transactions. And I’m going to say that perhaps only Bob and Group 1 are the appropriate groups to have access. Now, obviously, you’d have to give considerably more thought to who should have this permission than me just randomly clicking things in the demo. But the principle holds. So I’m just going to say, right, only those people should have access. And you see very quickly in this column here, we’ve now granted the admittedly questionable alter authority against transaction, but to a much smaller group of users. It’s no longer universal access. We’ve made a small change that makes this system that tiny bit more secure than it was. And my advice would be when you initially load up your RACF data into Security Discovery, have a good look at the to-do list, see the things that we call out and say, this could be better, and think, do I want to address this?

[00:26:25.] – John Taylor
Now you might say, no, I’ve got a perfectly good reason for Whatever it is. And then you do have the option of hiding any given issue and saying, nope, on my particular system, that’s not a problem at all. That’s just, that’s just fine. That’s the way we do things. It’s okay. And you can hide issues to make the list go down. But a good goal to aim for is to try to make this to-do list as small as possible over time. Now, I could spend quite a long time talking through the different to-do list items here, but there’s lots and lots of things in discovery that I want to show you. So I’m going to cheat a little. And I’m going to move to a completely separate piece of test data to show you another feature. And this, this is actually a feature that we introduced into the security discovery quite early on. And this is aimed at a specific problem that some customers talk to us about. And it’s the one that I mentioned earlier about how you grant authorities to individual users over time. And so what we have here, we can see that we’ve got a whole bunch of users, but we’ve got no roles defined.

[00:27:25.] – John Taylor
They’re just Everybody’s an individual in this system, and we’ve got lots and lots of different resources, lots of different transactions that some users have access to. And the problem is that we would really like to manage this better. We can’t confidently manage who has access to what, to what resources. We can’t confidently say, well, I can take this permission from this person when they leave this role because we’re not sure what the impact will be. Now, I can’t wave a magic wand and make it all instantly better for you, but I can help you quite a lot here. And the feature that I’m going to talk about now is we have the ability to analyze the data in the editor. Now, I will mention that the analyze process can take a while. It’s a k-means AI algorithm, so thankfully there’s no large language models involved, but it is an AI algorithm that has looked at your data and says, can I come up with sensible groupings of these users, of these resources? That’s a good step towards what you want. And we’ve got 3 types of analysis. I’m going to talk about 2 of them for a moment.

[00:28:36.] – John Taylor
Exact analysis is quite simple. We put users into the same role if they have the same permissions, and we put resources into the same member list if the same users access them. So you get nice firm groupings where Absolutely everybody in that grouping shares the same permissions. So life is simple and that’s a really useful starting point. Now for some users, a more useful one, and this is going to vary for you exactly which one of these you want. Similarity analysis is a bit more interesting. That’s the one where we say, okay, we put together users that have exactly the same access, but also we’ll have a look at the ones that seem to have a lot in common. So for example, you’ve got half a dozen people in a given role and they share lots and lots of permissions in common, but there’s one or two permissions that one or two of them don’t have. Now that’s either because there’s a really, really good reason they don’t have that permission, or it could be that they’ve just not been granted it yet. Somebody forgot to give them this permission. Now you’ll know in your organization which one of those sounds closer to where you want to start working.

[00:29:46.] – John Taylor
But so I can show you some of the features of the editor, I’m going to use similarity analysis and I’ll show you what happens. So fortunately, I’ve not got a lot of data here. This is going to run quite quickly. If you’ve got lots and lots of users, analysis can take a while. It can be an excellent time to go off and grab a coffee while it runs. But that’s going to run nice and quickly for me. And what you can see now, if I scroll around a bit, you’ll see that we’ve now got some nice fairly solid-looking groups of users. Now, I mentioned that there may be situations you’ve got a user who didn’t have a particular permission. So Sid Gregory’s here, here’s a really good example. We looked at these users and said, well, these users seem to be basically doing the same job, accessing the same transactions, but there’s one or two anomalies. And Sid doesn’t have yet, doesn’t yet have permission to transactions T002. Neither does Kevin Peterson. Incidentally, if any of you are fans of cricket, you might notice where we got all these names from, but that’s kind of an aside.

[00:30:48.] – John Taylor
It’s not that important. Now, there’s a couple of things we can do about the fact that we’ve proposed that maybe Sid should be in this role and should have this permission. Now, either you can say, I approve Sid being in this role. Yes, it’s right. Sid should have access to the transaction. It’s a mistake that we haven’t given it to him. And we probably would have got a phone call sometime next week from Sid saying, why can’t I perform this action that I’m meant to be doing as my day job? And that’s really, really simple. We just say approve that proposed change, and then Sid is part of this role with that permission, and it all looks nice and straightforward. Now, there’s still the question of what to do with Kevin and so on. Now, if you’ve gone down the similarity route rather than the exact route, if you go down the exact route, you haven’t got any pluses because everyone gets put where they need to be. And you might say, well, okay, I need to add 1 or 2 people into the roles. But equally, you might in this scenario say, no, Kevin really, really doesn’t belong in this role.

[00:31:45.] – John Taylor
He really should never have transaction T002 because I don’t trust him with it. So we could say, let’s move Kevin, this user, to a completely different, maybe to unresolved, and say we’ll figure him out later. Or we could say we want to approve it. We could even say— Kevin is really, really special, and I want to create a role based on Kevin’s permissions and then move other people into that role. There’s lots of things we can do here, but I’m going to do the simplest one. I’m just going to move Kevin right out of this group and say he’s no longer my problem. So I’ve got— I’m working slowly towards a group where there are no pluses. It’s basically what I’m trying to achieve. I’m trying to achieve a situation where we say everybody in this role has the right permissions. They’ve all got the same permissions because, say, if you’re in the same role, by definition you have the same permission because it’s the role that has the permissions, not the user. Now, I could go through doing that with all of the users, but I’m going to run out of time doing that.

[00:32:45.] – John Taylor
So what I’d like to do is I’d like to stop looking at this bit of data here, and I’m going to cheat in a minute, I’ll be honest. I’m going to show you a newer version of this data in a bit when I want to do the next stage. Well, I want to start looking at data from the CICS region. But as you can see, I would— what I would typically do if I was doing this for real, I’d go through looking at each of these users saying, is this user really, really in the right place? Now, if you did exact analysis, that’d be a much quicker process, but you’re still going to want to have a sanity check just to say, is anybody in a role they shouldn’t be in? Is anybody sitting here who gained these permissions 20 years ago when they worked in a different role? They should have lost them a long time ago when they became a manager, but nobody dared take the permissions away. Now’s an ideal opportunity to look at this and say, that looks really, really odd. Dean James is a manager. He shouldn’t have access to run these transactions.

[00:33:37.] – John Taylor
So now would be a great time to take those permissions away from Dean, put him in a more appropriate role where he can do less damage. You can see I’m quite positive towards management, can’t you? Put him in a role where he can do less damage and just leave him with the permissions that managers need and nothing else. So I’m going to stop demoing for a moment. I want to move on to the slides. There’s quite a lot of information I want to cover. And I’m never quite sure whether this is all going to take too long or run a bit short, but we’ll have plenty of time for questions at the end with any luck. So bring the slides back. I showed you a demo using just the data from ESM. So the fun bit now is going to be, let’s start looking at what data we can capture from your running CICS operations. Now, remember I mentioned this is, this is something running prediction. There’s quite a few points I need to mention about capturing security discovery data. Most important thing is you’ve got to be running CICS version 6.2 or later to capture this data.

[00:34:34.] – John Taylor
If you’re running 6.1 or 5.6, you can do everything I just showed you, add a bunch of other things, but you can’t yet do this bit. This is the bit that you need the newer CICS for. And what we do with security discovery data, once you switch it on in one or more CICS regions, we discover quite a lot of information about which resources are used by which users. We discover that firstly based on which is the user ID touching that resource, but also which transaction was being driven to touch that resource, and what’s the origin transaction that drove that transaction that drove that resource. So obviously, you have a chain of transactions. A caused B caused C, and C touched the file, and we keep track of that chain because that comes in quite handy later, as I’ll show you. A really important point to make is that when a user accesses a given resource, we just record that fact once. So we do not record the difference between somebody touched this resource once and touched it a million times. We merely capture the fact that it was touched. For the very simple reason that if this user’s touched this resource and you don’t grant them authority to capture it, they’re going to be pretty miffed about not being able to do their day job.

[00:35:53.] – John Taylor
Either that or they were hacking and shouldn’t have been touching it in the first place. But either way, the number of times they touch it is not important for this particular question. We capture data in 64-bit memory because there could be quite a lot of it and we don’t want to waste valuable storage. We expect security discovery data to be captured over a fairly long period of time. And I have a slide in a minute which will talk a bit more about that. There is a performance overhead capturing security discovery data. At the end of the day, nothing’s free, but we’ve worked very hard to make the performance overhead extremely low. It’s also an overhead that drops quite rapidly over time. And again, I’ll talk about that in a moment. You can switch on security discovery ideally by a simple POT program so that as soon as your CICS TS starts up, we capture absolutely everything that’s happening in the If that’s not appropriate for you, you have the option of switching on or off through CICS Explorer, even through the SPI if you wish. So for example, you might want to switch off security discovery when your system’s really, really heavily loaded, just when you want to eke out every bit of performance you can.

[00:37:05.] – John Taylor
One really interesting feature that’s something we can do because we’re running deep inside CICS. When we capture security access, we don’t just capture things like a check where you’re going— where, say, you’ve got file security switched on. We don’t only check things for the security checks you already do. We check the security checks that would have been done if you switched on that security setting. So if you’ve switched on file security, then yeah, we capture it. But if you haven’t switched on file security, but you’re thinking about it, we can still capture which file accesses occur, and that’s really, really handy if you’re trying to figure out what permissions you should grant to your users before you switch on security settings. We also ignore the RESSEC and COMMANDSEC settings in terms of capturing what security checks have been performed or would be performed, because that’s the key here. We’re not only checking the security checks that really went through RACF, we’re also doing the checks of What Check would go to RACF if you had that security switched on. We write the data daily to a log stream. Obviously, we can do it on request or at shutdown.

[00:38:17.] – John Taylor
And if you’re running this against multiple CICS regions, which is not unreasonable if you’ve got a selection of cloned regions, that data can be combined offline. So quite a lot of data on that slide, but some important points. Very quickly, I mentioned that there is an SPI for security discovery, so you can switch it on, you can switch it off, you can specify which classes of data you want to collect. So if you know, for example, that you— all your interesting security data is in files, that’s the thing you’re going to focus on recovering data for. If you have no interesting data at all in files or no particularly secure data in files, you might say, well, I don’t need to capture file accesses, that’s not a problem in my particular regions. So that’s something that Again, you can make your own decision about. Now, obviously, we collect stats, and one of the more interesting pieces of information is just how much data has been captured recently. And we can write it out with the SPI. Yes, that’s the next slide. Here we go. So we do expect security discovery to run for quite a long period of time.

[00:39:28.] – John Taylor
The reason being, if you ran it for just a day, You would certainly capture most of the file accesses, most of the access that your users perform. But there’s going to be special periods which you, you know, in your calendar when certain things happen. But, you know, end of month processing, end of year processing, that sort of thing. Really important to capture those edge cases so that you don’t say, well, nobody ever uses this file which is marked end of year data, so we don’t need to give anyone access to it. So It’s really, really important to capture all the interesting time periods as well so that you don’t miss out on the key bits of data. But a nice thing about the stats data is you’ll be able to see how much data has been collected, how the data tails off. So once we’ve seen a particular access happen, we don’t have to record it again because we’ve already seen it. So you’ll find that quite rapidly the stats The amount of data being collected tails off very quickly. The small overhead that I mentioned becomes even smaller because it’s data we’ve already captured and we don’t need to capture it.

[00:40:34.] – John Taylor
But that also gives you a fairly good clue of when you’ve captured enough data. When we stop capturing anything new, that could well be a good clue that you’ve captured everything that needs to be captured. So back to that picture again. You remember that Last time I highlighted this diagram, I highlighted the bit at the top with data being extracted from RACF. We’re now onto the SDD data being captured from CICS region. And as I said, this data gets captured to a log stream. We provide a really simple JCL job that extracts the data from the log stream, writes it out to a CSV file on USS. You’ve got an example of the SDD data down there at the bottom. It’s sort of kind of human readable, but certainly not as pretty as the YAML file I showed you earlier. Not really intended for human consumption, but the Security Discovery Editor is quite happy to read that data in and show you lots and lots of information once it’s been pulled in. In a moment, when I get to the second demo, I will start to show you the sorts of things that you can do with all that data and just how cool it can look.

[00:41:42.] – John Taylor
in the editor. So it looks a lot prettier than the CSV file. And yeah, as I said, got the data into Explorer, and now we start— we can start using that data in the Explorer. And this is mostly better to show you as a real demo, but there’s a couple key points I mentioned on this slide probably worth mentioning. The obvious one is when you can see what permissions users have actually used over the time you’ve been running the analysis, you can compare that against their permissions. And you can start asking yourself the question, why, why is this person granted this particular permission which they never ever use? And it might be that they, over the period of time you captured it, they didn’t have to run that relatively unusual function, or it might be that they’ve just got way more permissions than they need. And if we go back to the Zero Trust idea, if somebody’s got permission they do not need to do their job, they shouldn’t have that permission and we should take it away. And this gives us a lot more confidence looking at users and their permissions to say, I can really, really take that person permission away.

[00:42:46.] – John Taylor
They don’t need it. I thought they didn’t need it. Now I know they don’t need it. So it’s a nice way to start cleaning up your permissions and making sure people haven’t got permissions they shouldn’t have. Also, things we can do with this is we can have a good look at those groups of users, those member lists that we created earlier, or I created very briefly earlier and then skipped over most of it. We have a look at that and say, Are these the right groups? Are these groups too big or too small? Have we got resources in a given member list that really, really don’t fit? You start asking questions like that and try to refine your data in RACF better. The idea here being not just you’re improving your zero trust stance, you’re making sure that people only have the permissions they should have. You’re also ensuring you’ve got an easier to manage collection of users, easier to manage collection of resources. When you know with confidence that all the users in a given role are the only people who should be in that role, and you know that all these resources in a given member list are very directly related by the function they perform, then you can start to have a lot more confidence, be a lot sure about how you manage your users and how you manage your permissions.

[00:44:00.] – John Taylor
Generally, the intention is this will make your life easier going forward as you try to handle the huge number of users, resources you’ve got. You’re not spending quite so much time thinking, have I got the right permissions here? Do I need to take something away from somebody? Now, moving forward, a couple more slides before we get to the second demo. So far, we’ve just looked at transaction security. So like I said at the start, we have the presumption that you’ve got security switched on in production, you’ve got transaction security switched on, But there’s a fair chance that you haven’t yet got security switched on for some of your other resources, that you’d like to put security onto more resources, but you don’t know with certainty exactly which permissions you need to switch on. And that’s something we think we can help you with. Now, we think that most customers typically don’t have a full automated application suite that’s got proper, thorough security testing within it. You may or may not have a good coverage manual test, but running all those manual tests and checking the security against it again can be quite challenging. So deciding which roles you want to create for, particularly for your non-transaction resources, can be very, very difficult.

[00:45:16.] – John Taylor
It can be very hard to figure out exactly which permissions you want to grant without breaking your users doing their day jobs. And obviously you start switching on security and stopping people doing what they need to do, people are going to get fairly grumpy at you fairly quickly. So I’m hoping what I’m going to show you in a moment in the demo will show you how you can start to address those problems. So without further ado, I have another slide that mentions a demo. So 2 demos today. I’ve been kind of generous. So I’m going to cheat a little bit at this point. You remember I’ve got this list where I started organizing users into roles. I haven’t yet given them sensible names. I probably should have mentioned this earlier. When security discovery generates the roles for the users and generates the member lists, I can’t know what, what the correct name is for a given role. So all of these roles you see here I’ve created with lowercase letters purely to make it quite clear to you that these are generated names you can’t pull into RACF. And you’ll have to look at them and say, oh, I know Arthur, he’s part of the Sysprog team, or I don’t know, we’ve got down here Roland maybe is a manager, whatever.

[00:46:29.] – John Taylor
Now I’m assuming you all know that and you’ll be able to turn those into sensible names. And it’s easy enough to rename them, but I’m going to cheat and say, here’s one I did earlier. Now it’s probably worth mentioning that I have no imagination, and so I’ve got some really stupid names for some of the roles and some of the member lists. But obviously you’d give these far more sensible names, and you, at this point, you, you, you’re fairly confident you’ve got everybody in the right role, that you’ve got all your transactions into the correct member lists, and life is fairly good. You’re feeling fairly smug about things at this point, and you’re thinking, yep, this is a good layout, I can work with this. And this will be a fine point for you to say, I’m going to put this lot in RACF, and these are the definitions I’m going to use from now on. Because you’ve got the obvious advantage that if we take, for example, Paul here, for an example, and we move Paul out of admin and he becomes an HR person, then taking away all the permissions Paul’s currently got that he doesn’t need because he’s leaving admin and giving him just the ones he wants to do in HR.

[00:47:31.] – John Taylor
Quite challenging if they still had individually assigned permissions, but moving Paul from one role to another is either an edit in here or a couple of commands in RACF just to change the group he’s a member of. So as I say, management gets easier. But a couple of things I’d like to mention. The first one is I’m going to download the SDD data that we created earlier. Again, it’s fake test data, but something quite cool that I’d like to mention is that it’s a feature that we added fairly recently to the editor. So if you’ve got a version of Explorer you downloaded, I don’t know, say earlier this year or maybe last year, you won’t see some of these features. You need the newest version of Explorer to see this. We’ve added the ability to, rather than load the files that I’m using from your local workstation, we can load the files directly from z/OS. Now part of the intention here is for some customers, they’re a little wary of allowing their security data to leave the mainframe and to be stored on files off mainframe. So you have the option of saying, right, I’m going to load my ESM file from a USS file over on ZOS, and it won’t exist as a file on the local workstation.

[00:48:49.] – John Taylor
And more than that, once you’ve loaded the data from ZOS rather than from local workstation, I won’t let you save it to the local workstation. You’ll be forced to save whatever you save back to ZOS. So we’re making it quite hard for your data to leak out of your ZOS systems. Now it is still in memory in the CICS Explorer. Yep, I can’t change that fact, but it doesn’t get written to the file system on the local workstation. Now for the purpose of this demo, I’m just going to use local files because most of these got created manually on my workstation by one of my colleagues. So I’m just going to use local files, but I want to make it quite clear that If you want to, you can just work with files on ZOS and never have local files loaded up to your local machine. But once you decide you’re going down that path, we force you to stay on that path so somebody doesn’t pull the data down into the workstation and then save their local file system. Kind of defeats the object to try and keep it secure. So I’m just going to load a local file, and this is again fake data, but it’s fake data that’s sort of realistic, and this is pretending to be SDD data captured from a CICS region.

[00:50:05.] – John Taylor
Now a couple of things have happened that are worth mentioning. The first thing to mention is, if I just scroll that up a bit, there we go, you’ll see that right now I’ve gained a whole lot of lowercase r’s here and there, and if you look in the details panel, you’ll see that we’ve got an access of read, a permission of read, but we’ve also got an SDD access, so data capture from the CICS region of read. And because we’re good citizens, we’re writing nice code here and define it all correctly. We’ve got a UAC of none, which is good. Now, that’s all great, but then we look at Matthew Hogarth here and say, well, Matthew has permission to access transaction T20, but never, ever, ever uses it. It begs the question, should Matthew really, really be in this role? And the answer might be yes, it might not be, but it’s one to look at and say, I’m not sure about whether that— excuse me, drinking some coffee. I’m not sure whether that user really should have that permission. I’m not sure whether they need it. Sorry, very dry throat. So they’ve either— it’s right, they have that permission and you’re happy and they should be in that role, or you can either say, well, no, Matthew shouldn’t have that permission, at which point either You can take Matthew out of that role and put him somewhere appropriate, or you might decide you want to remove the entire transaction from the member list and say, nope, this member list was wrong.

[00:51:31.] – John Taylor
Again, that’s something that we can draw your attention to, but that’s something you’ll have to decide for yourself. You have to know for yourself. Now, I did mention that we’re capturing data for not just transactions, and for this particular type of test data, you can see there’s lots of different types of data we can capture. And that’s down to you when you switch on security discovery. Which of these do you want to capture for your particular CICS regions? Which things do you feel are really important to capture? Which things are you less interested in? And again, that’s something you will know for your CICS regions. I can’t tell you that, but you decide on which ones you want. And I’ve said, well, I’m interested in files. So if I look at files briefly, you’ll see that I’ve— if you look closely, you see I’ve got lots of familiar-looking user IDs. And I’ve got lots and lots of files listed here. I’ve got no transactions listed, quite importantly. Now my files are not in member list yet because we don’t know much about them, and you can’t actually see any Rs or any accesses yet. So I’ll scroll down a bit.

[00:52:33.] – John Taylor
Here we go. Under the unresolved section, you’ll see that we’ve got R+R, we’ve got U+U. So let me explain briefly what’s happening there. Every place that you see a lowercase r, We’ve seen this user read from this file at least once over the time period we were accessing, or in the case of the U, we’ve seen somebody do an update and so on. Now we’ve also got the capital letter R+ or the capital letter U+, which is quite simply a proposal that says, because we know this user accessed this file with this level of access at some point, we know that you need to give them the same access. as permission in RACF, or they won’t be able to do their day job. Or of course they’re hacking and they shouldn’t be touching the file, which is a different discussion to have with them. So at the moment we’re saying we’re proposing that you’re going to want to grant all these permissions. You probably are. And something we could do at this point to start again grouping these users into appropriate roles and grouping the files into appropriate member lists. But on most CICS regions, you’re going to have a lot of different applications running.

[00:53:46.] – John Taylor
Well, sorry, rephrase. Customers kind of fall into 2 camps. You either have CICS regions that are typically running one given application, that’s all that everybody does in that application, or you’re going to have lots of different applications in the CICS region. And I’m mostly going to talk for the moment about the second scenario where you’ve got several applications all running in the same CICS region, and it makes it very hard for you as a Sysprog for you and your colleagues in the, say, the application owners, etc., to look at the big picture because there’s so much data. So if I go back to transaction views, quite cool thing we can do with all this lovely data we extracted from CICS. I’m going to create an application filter. So there’s lots of definitions within CICS, within ZOS, of what is an application. And what I want to talk about here is The definitions of an application we’re talking about specifically for security discovery. And from our point of view, an application is a collection of transactions that are run by one from invoking one or more origin transactions that may or may not access certain other resources.

[00:54:54.] – John Taylor
And that sounds great. Let me give you a demo that makes it more sense. I’m going to create a filter that I want to look at just one application. Imaginatively, I’m going to call the application Demo, and you’ll notice that quite a lot suddenly changed. My table has gone completely blank, and I’m pleased to say that that’s not a bug in my code for once. That is quite deliberate because you’ve got an application filter that says only show me the information for the selected origin transactions, and I’ve not selected any origin transactions yet, so I’ve not said there’s anything I want to see. So I’m going to say that I have an application on my system, and I know there’s 2 ways of invoking this application. It’s— and that is by invoking transaction T002 or T006. Completely arbitrary, but suitable for my given test. And it looks a bit iffy when you see that the member lists associated with it are mortgage and credit. But I did say I wasn’t very imaginative on names of things. But I’ll also point out that over here we have a list of transactions. Now that list is all the transactions that get invoked because you invoke the origin transaction T002 or T006.

[00:56:07.] – John Taylor
Now obviously 2 and 6 are also included here because they are part of the application themselves. But I also know from the SDD information that when you invoke T002 or T006, any of these might get driven. So I’m going to say, well, my application therefore must contain everything that’s in mortgage because that’s 5 of my 7 transactions already included. And there’s this slightly dubious-looking 2 out of 7 for credit, which I’d like to show you in a moment. So the good bit, going back to the table, I’m going to make it a little bit wider so I can show you a bit more information. If you look at— where’s— there we go, mortgage. There we go, mortgage. If you look at mortgage, you’ll see we’ve got these several transactions that are all part of mortgage and say, yep, life is good, we like those. And users under role test all access it. Again, life’s nice and simple. But if I go back to credit momentarily, you’ll see we’ve got some warning triangles here. And what these warning triangles are telling you Is that while these transactions are included in this member list, because you said these are— these transactions are all grouped together logically, they’re all part of the same thing, the same application.

[00:57:25.] – John Taylor
When we looked at your running CICS regions, those transactions were never run as part of the related transactions. Now, that might mean that these are transactions that extremely rarely run. And you know that these should be associated with other transactions, or it may mean that your member list is just too broad, and you’ve included transactions in here that have nothing to do whatsoever with the credit application. Now, again, I can’t know that for sure. I’m just presenting the data that we captured and saying to you, this looks a bit odd. And again, it’s an opportunity for you to look at it and say, is that member list really, really correct, or are we missing something important? Have we made this too broad. And because that’s not a question I can answer right now, it’s not a question I quite want to answer at this point. I’m going to cheat a little, ignore the problem completely, and go back to the files, because files are interesting. I mentioned earlier that we capture data for lots and lots of files, and I showed you there’s a nice big table, lots and lots of information in it. Now, depending on how good your memory is, You might see that we’ve got far fewer users listed here in unresolved now.

[00:58:37.] – John Taylor
And we’ve got— there’s not so many files to worry about. And the reason being that we filtered this down to just the files that have been accessed as part of driving those origin transactions. So remember I talked earlier about the thought of, well, I want to switch on file security, but I don’t know what permission to give to people. Well, first of all, I’ve shrunk the list quite a lot down to just the ones involved in this application. But I can go one stage further. And I mentioned earlier that we’ve got the analysis. We’ve got 3 types of analysis. I talked about exact analysis, which I quite like. I showed you similarity analysis because it gave me an opportunity to show you the whole question of should this user really be here. But I completely ignored best fit analysis. For the very good reason that you can only run best fit analysis against things other than transactions. So best fit analysis is specifically aimed at the scenario where you don’t have definitions in RACF yet, where you’re trying to figure out what the definition should be in RACF ready for switching on that type of security, security against that particular resource.

[00:59:47.] – John Taylor
So now would be a really good time for me to run best fit analysis, having talked about it. And again, things have changed, and I’m going to scroll up a bit too because I want to show you something moderately interesting. Way back, we created the incredibly questionably named role called Others, and that was a role specifically based on transaction data. Now, when we looked at the file accesses and we looked at which transaction drove which files, We were able to say, okay, these file accesses— you’ll notice the naming here, othersxfct— this is all about file accesses that are in some way related to that role called others. And we further split it into 2 groups. We’ve got xfct or xfctu. Now, depending on how you want to manage your roles, and different people do this different ways, you might say, My preferred option is to have completely separate roles defined for transaction access and for file access. And obviously, if you’ve got different levels of permission, again, you’re going to need different roles. So I’ve created 2 completely separate but closely related roles for you. I mean, the XFCT one’s not very interesting because there’s only one user.

[01:01:05.] – John Taylor
See if I can find one with more users. Have I got one better? There we go. Test is way more fun. So if we look at test, you’ll see that I’ve got several users in test, and I’ve got these 2 bigger roles here. Extremely unlikely looking data because it would appear that we’ve got more people who can update files than can read them. But, you know, it’s test data at the end of the day. So as I was saying, you might want to keep these as 3 completely separate roles in RACF, and that’s completely fine. Or you might say, I mean, it doesn’t quite fit this data here, but you quite often, you might say, well, everybody in role test also needs to read the file from test users. So you might want to, for example, merge this role testxfct into the main role test. Or you might not want to. You might want to merge as intersecting one. Intersecting one, sorry. Of course, we’ve got the thing about these are all pluses at this point. So you might say, yes, I’ve looked at that, I’m happy with those users being in that role, and I can approve this and make all those pluses go away just like that.

[01:02:14.] – John Taylor
I’ll do that one just because I should have mentioned it earlier. But I think the key point I’m trying to make here is that you’ve got the ability to say, I want to manage these as 3 separate roles, the transactions, files of read, files of updates, or you can start merging them and say, well, if you’ve got access transaction, need read, and manage them in a single role and then keep the updates separately. That one is again your choice. I probably should have mentioned earlier along this little line here, as we’re working, different bits of data change here. We tell you at any one time what the model is working. That model name is based on the file, the ESM file loaded originally. Right now I’m looking at files. I filtered it down just to an application filter that I called demo. And then this is how many roles and member lists being displayed. But there’s also the question, well, how many things are being hidden? And because we’re in this filter, because there’s lots of stuff that isn’t related to this application, there’s 38 resources not being shown at all. Now there’s a bunch of other icons I haven’t talked about yet, some more interesting than others.

[01:03:24.] – John Taylor
They’ll do collapse and expand. You can create member lists and things. You can create roles manually. I really, really should have shown you the compare feature that I talked about earlier. So I’m going to say, because the simplest one, I want to compare this model I have now with the data I loaded up originally. So when I loaded up this data way back, probably forgotten that far back now, we had a bunch of users that weren’t enrolled at all. We had a bunch of resources that weren’t in member lists. So I’ve changed quite a lot. Oh, and I’ve loaded up a whole load of data from the CICS region, so this could make for a fairly interesting or busy-looking report. Let’s have a go, should be fun. I’m going to export this to somewhere that I can find it easily, just for convenience. Now I’ll put in Downloads, and I’m just going to call it Compare. Okay, and so it’s going to get put in Compare, and it’s being written out. Bear with me a moment while I fire up another GIX Explorer. Not CICS Explorer, Windows Explorer. Remember which explorers I’m doing. And in Downloads, I have a file.

[01:04:34.] – John Taylor
It’s just a text file, so I’m just going to open it up in Notepad, nothing clever. And very quickly, and you can do this report as often as you want, it can tell you what sort of things you’ve done. Now, unsurprisingly, I’ve got loads and loads of resources in my model that weren’t in the original ESM file. for the very simple reason that we didn’t have any file permissions in RACF because we hadn’t switched on file security in CICS. So there’s loads and loads of resources listed there. We started off with no roles at all, so we’ve added a bunch of roles. We’ve— yeah, here you can see the roles we’ve added. We’ve added lots of file member lists. And one thing I should have mentioned is because I’m filtered down to an application filter right now, this report only tells me the information for this application. And that’s a very simple reason that if you’ve got several applications running in a given CICS region and you’ve got a different application owner in your company for each application, you might want to show them just this data and say, this is the— these are the permissions that I think I should grant to people to use your application.

[01:05:43.] – John Taylor
Do you agree? Now, if I’d done— if I’d not got the application field switched on, I’d just see absolutely everything in the model that’s been changed. Much bigger report, but much harder for the application owner to have a good look at and say, does that look right? And as I said, lots of things added. So the nice thing here is, here’s a quick introduction to all the users that have been added to it, to a role. Because as I mentioned earlier, we created roles. So let’s go and look at What roles they’ve been added to. So again, it’s keeping track of things. Really, really useful section here that the application owner is probably most interested in is what new permissions have you granted in RACF. And you’ll see that there’s lots of pluses here which show that these are still subject to approval. But these are permissions that we’re suggesting putting into RACF that these users currently don’t have. And hopefully the application owner will say, yes, that looks completely sensible for those users to access those resources. One of the bits that probably make you a lot happier most of the time is saying nobody has lost any permission to anything.

[01:06:52.] – John Taylor
So with all that messing around that we’ve done, we’ve not accidentally taken away any permissions from somebody that they currently have today. We’ve not stopped anyone doing their day job. And that’s a good thing if you’re happy that nobody had permissions they shouldn’t have had. If there’s permissions that people should have had— sorry, the permissions that users did have that they shouldn’t, that’s the zero trust concept, then you’ll have items listed in there where you said, no, I’ve taken that away. That person shouldn’t have had that. It should have been taken away from them years ago, so I’m taking it away. And this is a nice confirmation for you that you’re doing that. And it’s a nice last-minute check. Am I confident that person doesn’t need that permission? So that’s that report which I briefly wanted to show you. There were— how are we doing for time? Yeah, I’m running a bit towards the end of time. There’s lots of other things in the Security Discovery Editor I could show you, but I have a few slides I’d like to get through before we take any further questions, and I just don’t want to miss things.

[01:07:54.] – John Taylor
So let me bring the slides back up briefly and I’m just drinking another mouthful of coffee as well. Sorry. Okay, so next slide. So, okay, very briefly, a summary of what we’ve looked at so far. Lovely big picture. The steps were extract your data from RACF using the JCL and REX, and that’s the bit you can do against any version of Git you’ve got. And you can have a good look at your data, you can look at it in the editor and start saying, right, What do I need to improve here? What, what, what is not great in my current RACF definitions? And you can work through that lot, improve your RACF definitions. And once you’re happy with that, that can be a really good time to say, right, life is good. Excuse me, I’ve got better, more manageable RACF definitions. Excuse me, I’m going to export these and I’m going to put them back into RACF and I’m going to start using them for real in my CICS regions. And life is better. Now, depending on what version of CICS you’re running right now, you could simultaneously be capturing the SDD data in your running CICS regions to find out what people really, really use and capture all this lovely information about the relationships between your transactions and transaction resources and to understand your applications better.

[01:09:13.] – John Taylor
And also spot things like which users have too much permissions, who has permissions they shouldn’t have. That sort of thing. So we’re capturing all that data, and again, once we’ve got the data into the editor, we can perform things like best-fit analysis to start creating definitions for other resource types. We can tidy up the definitions we’ve got already. We can check all the definitions that we thought were great earlier. And again, we can do an export whenever we feel like it, when we know that we’ve got better data. And again, we’ve got the opportunity to switch over to using these better definitions in your CICS region. And at some point you’ll go around this and you’ll say, right, that’s it, my permissions are great. And we do have other processes that can help you maintain that stuff and keep it good in running CICS regions. But that’s a presentation for another day. So the basic goal here is improving your definitions in RACF, making them easier for you to manage, and ensuring that you’re confident that they are easier to manage and better. So that was very quickly the steps we went through. The demos.

[01:10:18.] – John Taylor
I think these numbers are almost certainly out of date because I try to keep the slides up to date. The bottom line is always try to grab the latest version of CICS Explorer you can to use the security editor because I keep adding features to it whenever I get time. The first demo just used information from RACF. You can use it with any CICS version you want. And the JCL is available with CICS version 6.1 onwards. We do update that JCL and REX every now and again, and when we update it, we keep it in sync on all the CICS regions. So they’ve all got the same version of the REX. There was a fairly important update we did to that REX a little while ago. So if you’re running on a really old version of that REX, you’ve not grabbed any APARS, you might find that you need to update that just to get the best results. Demo 2 needs CICS 6.2 or later to capture the really cool information from the SDD file. As the nice thing about the fact I’m giving you these slides, you’ve got all this stuff to refer to later.

[01:11:25.] – John Taylor
So when we’re looking at the SM data, we loaded the data into the editor, but already you’ve got a table that makes it easier to understand your data than just trying to understand the output, lots of RACF commands. So before you even start editing anything, I would hope you’d agree that this makes things easier to understand, to understand what you have today before you start improving it. The to-do list I’m really quite proud of. That’s the bit that says these are things that we think you might want to look at, things we think you might want to improve. And I do add to this to-do list when I can find time and So each time you grab a new Explorer, I hope you’ll see something new on here if it’s relevant to your particular environment. I showed you the analysis, so the converting from a complete mess of permissions against individual users to more sensible groupings of roles and member lists. I showed you that when we load the SDD information, We can then find out a lot more about what’s really, really been happening in your CICS region, and there’s lots of things become possible then, including spotting who has permissions they don’t need.

[01:12:37.] – John Taylor
The application filter is an optional thing. You don’t have to define an application filter if you don’t want to. If you can get your head around all of the stuff running in your CICS region at one time, that’s brilliant. You don’t need application filters at all. If you do need to slice up your Git region into several different applications to understand each one, one at a time, that’s what the application filter is for. So that, that’s a really handy feature if you need it. And if you don’t need it, you can get by just fine without it if you’ve got a relatively straightforward Git region that just does one thing. I talked about creating security definition for files and using best fit. I did that Having defined an application filter first, but it’s entirely possible if you’ve got an appropriate CICS region to say, I’m going to do security definitions for everything on my CICS region at once without using application filter. And sometimes that makes more sense depending on just how complex your CICS region is and how many disparate applications you’ve got running on there. Summary. Okay, yeah. Fair point to start with.

[01:13:42.] – John Taylor
Zero trust is a journey. It’s not a destination. It’s a journey. It’s a mindset, and it can be difficult. A good starting point is knowing what you currently have, knowing exactly what you really, really have in your systems, and knowing with confidence what you can safely change without breaking people doing their day jobs, without breaking key applications, because people get really grumpy when you stop their ability to do their day job or when you break an application. So being confident about what you can safely change. It’s based on knowing what you have and understanding it. And I hope you’ll agree that security discovery helps you understand that information. The ESM and SDD data can help you a lot. Refine the ESM data before you even get as far as the SDD data gets you to a more manageable view of the model. It gets you to a position where you’re starting to think about roles rather than individual users, where you’ve addressed some of the more obvious anomalies that built up over time in your systems. And pretty much all of us have got dodgy-looking definitions somewhere, maybe a bit, bit too much UAC, or we’ve created 3 roles over time that basically do exactly the same thing, but we didn’t dare get rid of 2 of them because we weren’t quite sure who was in which one.

[01:14:56.] – John Taylor
That’s the sort of thing, again, the discovery adds to make it much easier to spot, and you can be much more confident about fixing those. So Possibly my favorite slide of the whole lot. Give it a try. Everything I’ve shown you, or most of what I’ve shown you, is actually really, really easy to try for yourself on your own system. So basically just 4 steps. Capture the ESM data from your own RACF. So run the JCL and the REX, which, as I said, you might need to ask the security admin to run. But I’m hoping your security admin will be okay with that. They’ll be able to look at the REX and say, yes, this REX is harmless. We haven’t got a problem with it. Once you’ve run that, load the SM data into your CICS Explorer, and depending on how your Sysprog feels about such things, how your security admin feels, you might be required to load that data directly from USS, or they might allow you to FTP the files over to your workstation. Arguably, it’s easier to just load it directly from USS anyway, just using the built-in connections in the CICS Explorer.

[01:15:56.] – John Taylor
Once you’ve got the data there, Have a look at your to-do list, and if you’ve got a completely empty to-do list where everything looked great, you can just be really smug. Tell all your mates, “Hey, I loaded up security discovery and I had no to-do list items at all.” You can be a happy bunny and forget about this because you’ve already done a great job of sorting things out. Or if there’s some items in the to-do list, have a good look at what’s there. Say, “Can we improve this? Are there things that we want to improve here that will make our life easier? Make get the auditor’s offer back, make the day job easier managing.” Who can perform what, make it easier to spot anybody attacking our system, or for that matter, make it harder for people to attack. And we’d love to hear what you’ve found. I believe Amanda can share my email address, can share Mark’s email address. We’d love to hear your opinions on what security discovery does. We’d love to hear how we got on with it. If you’ve got any problems with it, we’d love to make it better. Maybe there’s things that you’d like it to do that it doesn’t currently do.

[01:17:00.] – John Taylor
Now, if you want, we do have a feature in the JCL that allows you to anonymize all your data. It’s primarily for service reasons. So if you have— if you need to raise a PMR against us, something’s not working, you can send all your data in an anonymized form to us. But it also can be useful if you just want to say, hey, This looks a bit odd, or what do you think about this? Or maybe share a screenshot of your data if you’re ready to, if you want. We’d love to hear what you think about it. And I’ve only got one more slide, so I might as well bring it up. Thank you for listening. Thank you for putting up with me talking so long. I don’t know if anyone’s got any questions, any comments, any plaudits are always nice. I can’t see the chat window, I’m afraid. Amanda, if you can help me on that one.

[01:17:54.] – Amanda Hendley
I am seeing no questions coming in. If anyone wants to share, you can drop. I am seeing some clapping hands going across the screen. So good job.

[01:18:05.] – John Taylor
Thank you.

[01:18:05.] – Amanda Hendley
Yes. I’m gonna

[01:18:06.] – John Taylor
drink a little more coffee. I need—

[01:18:07.] – Amanda Hendley
Does anyone have any questions that they want to drop into chat?

[01:18:18.] – John Taylor
I’ll answer the most obvious question. Yes, it’s black coffee and it’s really nice.

[01:18:24.] – Amanda Hendley
Wait, what time is it for you that you’re having a cup of black coffee?

[01:18:26.] – John Taylor
What time? I don’t know. Um,

[01:18:29.] – Amanda Hendley
6 o’clock?

[01:18:31.] – John Taylor
10 to 6.

[01:18:32.] – Amanda Hendley
All right. Tea time. Braver than I am. Um,

[01:18:39.] – John Taylor
evening 10 to 6, I should point out, not morning.

[01:18:43.] – Amanda Hendley
Well, John, thank you so much for your presentation today. I don’t see any more questions coming in, but to answer the question we always get continually is yes, John has offered up the slides, so we will make sure that we can share those. Oh, there is a question coming in. Can I read it to you?

[01:19:05.] – John Taylor
Yeah, please, fire away.

[01:19:06.] – Amanda Hendley
Okay. Does Discovery Security Does the security work also with CA Top Secret or only RACF?

[01:19:14.] – John Taylor
The short answer is only RACF at the moment. The longer answer is we’ve had chats with the providers of the other security products, Top Secret, etc., and they’ve been interested. They’ve listened to what we’ve said. At the moment, we have architected it in such a way that we can work with their products, with their help. So as a minimum, they’d have to create their own version of the REX that I mentioned to extract the data from their ESM, convert it into our ESM file. But in theory, it should work. I suspect I will have to make a couple of code changes when they provide me with the data. I don’t think it’ll be an instant zero-effort challenge for me. But when they say to me, hey, we’d love to work with you, this is the sort of data we’re generating for our ESM, I would love to get them on board and get their data working. So that’s my honest answer. Okay, it’s the long answer. The short answer is right now only RACF because I’ve got quite a lot of access to RACF.

[01:20:20.] – Amanda Hendley
Sure. All right. Additional accolades coming in. So I have thrown up the 2 QR codes for our User Group LinkedIn where you can start a conversation or connect with other people in this space, and also our Mainframe Virtual User Group YouTube where we post videos. We also post them on our site. So if you will give us about a week, we’ll have this video and the presentation and a recap all ready to go for you.

[01:20:52.] – John Taylor
And now we’re YouTube famous.

[01:20:55.] – Amanda Hendley
You’ll be YouTube famous. That’s right. And next thing you know, you’ll have a YouTube channel of your own and maybe some TikTok dances.

[01:21:05.] – John Taylor
No, I don’t think I can see that one.

[01:21:07.] – Amanda Hendley
No. Okay.

[01:21:08.] – John Taylor
I think my son would be distinctly unimpressed if I started doing TikTok dances.

[01:21:13.] – Amanda Hendley
I mean, I can only think of one other person in the mainframe space who might be doing TikTok dances.

[01:21:21.] – John Taylor
And that’s Mark Pocock.

[01:21:23.] – Amanda Hendley
Oh, oh, not who I was thinking of, but okay.

[01:21:28.] – John Taylor
Mark’s about as likely as me, I think.

[01:21:31.] – Amanda Hendley
Okay, well, great. Well, are you gonna be at GSUK in a couple weeks?

[01:21:37.] – John Taylor
Yes, I’m presenting. I won’t be doing this demo there. I will be presenting what’s new in the CICS TS open beta. So all the stuff we’ve been working on, And I’ll confess, because it’s me demoing, I have slipped a slide in about security discovery because I couldn’t resist. But basically, it’s about the open beta, all the cool new features we put in. And there’s some good stuff there.

[01:22:01.] – Amanda Hendley
Awesome. Awesome. Great. Well, I will see you there. And hopefully, we’ll see a couple of our attendees today there as well. And with that, we’re done for the day. It was so great to see everyone. Again, you’ll get an email with all of this information from today’s session, and I will make sure that you guys get the invitation first for our rescheduled presentation from earlier this month. So thank you again. Thank you, John. It’s great seeing you all.

Virtual CICS Sponsor

Broadcom
Claire Connor

John Taylor

IBM Hursley

Upcoming Virtual CICS Session

Get the latest Virtual User Group Updates

This field is for validation purposes and should be left unchanged.
Select the Groups you'd like to Join(Required)