Data centre managers and IT operations staff in large enterprises are now faced with a bewildering array of demands from their users, who take for granted that systems will work without any form of service interruption, planned or otherwise.
At the same time the business expectations are for new services to be delivered rapidly and at minimal cost.
As the range of services expand and as workloads can vary unpredictably in the resources required to keep them working at the desired service level, senior IT staff are facing many operational challenges. What options are available to them going forwards?
We know that senior IT professionals are challenged by users to improve the perennials of reliability and availability whilst security and scalability are never far behind in terms of business concerns. These create problems with many of the x86-based server platforms as getting such characteristics baked into solution architectures can involve adding considerable work to already stretched IT professionals as management complexity increases.
As a consequence many organisations are looking at how they can deliver robust services cost effectively, with enhanced management and resilience, without placing additional stress on IT administrators and operators. It is interesting to note that the service characteristics being sought to support these business workloads match neatly onto a server platform that is frequently overlooked in organisations, namely the mainframe, of which the IBM System z is the most widely deployed platform today.
Many organisations have consolidated the x86 platforms they run onto a smaller number of larger machines. To date, most effort has been expended on consolidating machines running Microsoft Windows plus a range of Linux and Unix applications. As such projects have progressed beyond initial trials and into mainstream production, evidence is growing that the operational management of consolidated platforms is posing problems. The answer to most operational problems, and indeed repetitive, routine tasks, is still to “throw people at it”, despite this being expensive and difficult to scale.
It is understandable that few organisations have looked at making wider use of their mainframe platforms to support such workloads as many IT professionals are coming to such projects with little knowledge of the System z and its characteristics, even in organisations where the mainframe is a long established part of the furniture. The availability of so-called “offload engines” to run Linux and Java workloads without the need for additional IBM platform licenses is often overlooked, despite these solutions offering considerable management and software cost benefits.
Perhaps of greater import is that in many respects the mainframe is the most manageable and secure server platform available, but few IT professionals are aware of these attributes. For example, it is often cited that it takes fewer administrators to run workloads over extended periods of time on the mainframe than on any other server architecture.
When combined with the low energy and space footprint of such machines compared with distributed server solutions, it is perhaps difficult to see why greater use is not made of their ability to support, securely and robustly, a wide array of application workloads. Organisations who are actively extending the workloads run on mainframes report that whilst there is a perception that such systems are difficult to manage and expensive to acquire and operate, this is today far from reality.
There is a clear potential in many organisations for mainframes to run a wider range of applications cost effectively and securely. This capability is being expanded to offer similar capabilities not only to Linux and Java applications using the offload engines, but before very long will see standard x86 workloads running on dedicated blades managed directly by the mainframe’s workload management software.
Mainframe administration certainly requires system managers to develop new skills, but management software solutions are developing rapidly to assist in the routine operation of the platform. The mainframe has always enjoyed excellent centralised resource management and reporting capabilities, and has the longest history of successful operation of virtualisation solutions, making it an excellent platform to host highly variable workloads.
A little research shows the advantages of making wider use of the mainframe to be evident, if sometimes difficult to assign monetary values to. But the benefits are not widely recognised by people apart from those with experience of the platform. For those looking at a virtualised environment, the mainframe deserves to be considered alongside big x86 boxes, and it will be interesting to see how much traction the mainframe can get in the new era of the private cloud.
That said, whilst the mainframe is under-exploited, it would be a mistake to think that everything could run on the platform. For a wide range of valid business and operational reasons, including software availability, integration, proximity, IT skills and preservation of investment reasons, it will still often make sense to stick things on Wintel servers or other systems. The real message is to think about what’s running where and why. If this is done, many mainframe users will almost certainly end up making more use of the platform.
Tony is an IT operations guru. As an ex-IT manager with an insatiable thirst for knowledge, his extensive vendor briefing agenda makes him one of the most well informed analysts in the industry, particularly on the diversity of solutions and approaches available to tackle key operational requirements. If you are a vendor talking about a new offering, be very careful about describing it to Tony as ‘unique’, because if it isn’t, he’ll probably know.