The acronym “SaaS” first appeared in 2001 as a revolutionary concept. Now, Gartner predicts the SaaS market will reach $116B in 2020. As SaaS concepts are brought to the mainframe, what will the market look like in another 20 years? What if you could dynamically and automatically redirect individual mainframe workloads to the cloud based on business requirements, capacity needs and cloud suitability?
This is a significant challenge for the future. However, the benefits could create new opportunities for users, service providers, mainframe ISVs and IBM. I believe we can start by redirecting batch and less data intensive applications—like compilers— to the cloud.
The Foundations of Cloud Computing
Cloud computing dates back to the introduction of the mainframe. During the 1950s and ’60s, the idea of accessing shared computing facilities over secure networks was widely adopted for research and educational applications.
What we now know as cloud computing was commercially popularized about 15 years ago by Amazon, starting with Elastic Compute Cloud (EC2) and growing into a range of cloud services under Amazon Web Services (AWS).
Today, cloud-hosted web services are ubiquitous, with competitive offerings available from a variety of vendors. Businesses have immediate access to various virtualized computing services, from Infrastructure as a Service (to access components traditionally present in an on-premise data center, including servers, storage and networking hardware) to Platform as a Service (which adds OS and application development environments) and Software as a Service (for delivery of full-function applications).
These virtualized computing services revolutionized the infrastructure of business, delivering flexible pricing models, the ability to scale on demand and accelerated time to market.
A Market Opportunity?
Whereas we are fully immersed in the revolution of the cloud, the benefits of cloud computing have been limited to distributed computing and have eluded mainframe environments and
applications. This gap in the market—the ability to dynamically use cloud resources to execute z/OS* workloads—presents a tremendous opportunity in mainframe computing.
The mainframe remains the most secure, reliable and available platform in the market today, trusted to run the most sensitive, mission-critical and regulated workloads and data in the world. Yet mainframe users continue to fight the perception that the mainframe is expensive, inflexible, outdated and destined for extinction.
Without some of the analogous cloud technologies of the distributed world, mainframe leaders are faced with the same old choices for cost reduction and operational efficiencies. Some hire employees or consultants to manually move data and/or workloads off the mainframe, migrate to a managed service, outsource, or attempt to re-platform applications. This creates substantial challenges in time, cost and risk to complete and if completed, prove challenging to retain or regain control of systems, data, workloads and costs, not to mention the comparative compromises to security and availability.
Cloud Benefits on IBM Z
Imagine the benefits of cloud computing realized on IBM Z: similar pricing flexibility, scalability on demand and accelerated time to market. These capabilities would allow mainframe users to maintain their mission-critical IBM Z* workloads on their internal mainframes as appropriate for their business, but with new flexibility to dynamically select when, where and how to automatically and securely move individual IBM Z workloads to private, public or hybrid clouds.
While today an entire LPAR, CPU or data center can move to the cloud, I believe a new Z Workload as a Service (ZWaaS) capability would be key in the mainframe world, with an environment as a Software Defined Mainframe (SDMF). In this model, a granular unit of mainframe workload would be selected, based upon a policy or rule set, and be dynamically dispatched from a known and configured source mainframe for execution in an environment hosted by a service provider. Depending upon the application and the unit of work, some or all of the environment and data required for program execution would be dispatched along with the unit of work, and the results returned to the source environment.
For large mainframe users, the opportunity to easily shift workloads around to various clouds to achieve lower costs, fulfill excess capacity needs, and flexibly test and deploy new applications is key.
For smaller mainframe users, envision mainframe applications executing in the cloud, still controlled by the user but managed by a service provider—just like businesses use AWS or other cloud providers today to remove infrastructure in distributed environments.
A Revolution in the Making
ZWaaS would present a compelling opportunity for service providers to bring a variety of clouds to market, and for ISVs to create new tools to manage and optimize mainframe environments.
Given today’s high-volume, transaction-based world, with these cloud and other innovations, we believe the mainframe will continue to serve as the most secure, reliable and available platform in the industry. It’s time for the concepts of cloud computing in the distributed world to come full circle, back to where they originated so many years ago—and revolutionize the mainframe industry.