MaaS – Mainframe as a Service
Analyst Coverage: David Norfolk
With Mainframe as a Service – MaaS – mainframe (or “Enterprise Server”) MIPS are accessed via a Cloud interface, just like any other Cloud service. Since Cloud is fundamentally location agnostic, the mainframe technology involved may exist in-house (no need to “rip & replace”) or, in future evolutions, come from an external service provider behind an SLA.
Why is it important?
MaaS adds agility to the traditional mainframe model, which is supremely cost effective and which runs 70% of the world’s mission-critical global business transactions. Nevertheless, the mainframe is only cost-effective if the mainframe is highly utilised and this can constrain its agility.
With MaaS you need only pay for the capacity you use, without over-provisioning “just in case”, and can exploit small niches of mainframe functionality (e.g., hardware encryption) while an application is starting on its growth curve. A leading education company has used MaaS to align its IT strategy with the consumption-based managed services model adopted by the business. You are also not locked into potentially exploitative mainframe software pricing models, and this makes small mainframe deployments feasible. For example, cloud-based backup via RSM Partners is protecting a client’s small but business-critical mainframe, providing assured resilience and peace of mind.
The use of MaaS means that you don’t need to compromise governance in the interests of agility – a mortgage company has used MaaS to achieve compliance as well as facilitating business growth. And, the availability of both a CapEx and an OpEx funding model on the largest, fully utilised, shared mainframes allows cost optimisation without compromising performance or agility. A steelmaker has used MaaS to reduce operating costs by 40% over 5 years – while improving system performance by 92%.
MaaS assists the rapid modernisation of Mainframes and IT infrastructure without impacting existing service levels, using Mutable self-service catalogues, including hardware-assisted services. A semiconductor manufacturer has saved 20% of its current budget whilst assuring future scalability, using MaaS.
MaaS potentially simplifies and speeds procurement processes, facilitating mutable change, as you can acquire what you actually need, not what you might need, and implement simplified maintenance contracts. It facilitates agile DevOps-style operations and deployment, with transparent SLAs for performance governance and accountability, a single point of contact for infrastructure services and streamlined service delivery replacing Operations silos. It integrates the Mainframe (on or off premises) into an organisational Cloud architecture which allows any existing mainframe to be integrated with a cloud service platform and supports legacy systems but also promotes OpenSource opportunities, thus delivering flexibility for moving mainframe management on or off premises, thus reducing the risk associated with the possible loss of in-house mainframe expertise.
How does it work?
MaaS is different from most distributed systems PaaS or IaaS models, because a fully utilised mainframe is extremely cost effective (unlike a typical AWS server, say) and easily runs at around 100% utilisation with better than true “5-nines” availability. A leading travel operator used MaaS to increase both server utilisation and peak transaction volumes by 20%, whilst improving resilience and time-to-recovery after disaster.
However, a mainframe has a high cost of acquisition, can be expensive to upgrade when it reaches capacity, and may require specialised support. MaaS offers, potentially, the simplest solution to these issues. For now, you exploit your existing mainframe; when you need more capacity, you rent it as a service from a mainframe specialist who manages and supports a large scale mainframe data-centre, with the associated support expertise (and possibly, mainframe training programs).
Note also that the new availability of commodity languages such as Java, and new techniques such as DevOps, mainframes has opened them up for general use.
A mainframe is simply the highest performing, most reliable, and most cost-effective source of computing power available, “just another computer” (albeit a very effective one). This is why some 70% of mission critical transactions still run, in part, on mainframes today. The issues with mainframes are: Initial acquisition cost; the availability of support staff; discontinuity as you upsize from a small mainframe to a bigger one; and, lock-in to a limited set of mainframe suppliers.
All of these issues are addressed in the hybrid cloud MaaS model.
“IDC and CA in their Connected Mainframe videographic, see MaaS in the “early adopter” phase and, thus, as the next opportunity for gaining business value from mainframe modernisation.”
“Your tools have to support any type of transaction, wherever… parts [of it] run. So if you are serious about demonstrating that your mainframe is just another computer, make sure you have the solution to prove that.” Marcel den Hartog, Sr Director, Field Marketing, CA Technologies
A potential MaaS customer needs to review its mainframe architecture, identifying services, required SLAs and Cloud-style interfaces. It then needs to apply a logical Cloud architecture on top of its existing mainframe status quo – with hybrid cloud, applications can even remain on-site. More than this, when workloads change or new applications are needed, utilising an off-premises mainframe-based service solution is a new design option available, if the move makes business sense. “On the Cloud, no-one knows that you’re a mainframe” – except that you can have more trust in any mainframe-based SLAs you are offered. The fundamental benefit of MaaS is that you are isolated from evolving technology (you can even use distributed 8086 if/when it can offer mainframe-type SLAs, if your applications can exploit distributed systems) and can take advantage of whatever infrastructure technology is currently most cost effective.
The Bottom Line
Mutable modernisation initiatives mustn’t compromise the legacy systems and existing customers who are paying for the modernisation – “reputation risk” associated with legacy customers can compromise the success of Mutability. MaaS is just the simplest and most effective way to modernise mainframe legacy and to make mainframe levels of performance, concurrency, security and reliability available to everyone, as a basis for the Mutable Business – “all the mainframe benefits you are used to at much lower TCO and with greater flexibility and agility”.
Further resources to broaden your knowledge:
The mutable mainframe
Despite appearances mainframes can be agile and can participate in DevOps and continuous delivery.
The Mutable Mainframe - the Mainframe is still very much part of the CA Technologies roadmap
The Mutable Business probably doesn't care much about what is behind its SLAs, as long as they're met. So, it might as well be a Mainframe.
So, what do I really think of the mainframe today?
Mainframe: Alive and Well!
A BMC survey that should be read by everyone, not just mainframe specialists
Enterprise server 3.0 – zEnterprise BC12
A new z mainframe arrives - and I wonder why it still seems to be a specialist choice for many companies
The Web + mainframe + Itanium = Very Interesting
Mainframe systems and applications could prove key to expanding transaction-based services via the web, and could provide a boost for Intel's Itanium processor as well.
Service Oriented Legacy Architecture – MF meets SOA
The mainframe of many companies holds critical services developed by talented and experienced COBOL programmers. SOLA provides an easy way to make these services an integral part of an SOA solution.