With the advent of IBM PureSystems, many clients ask themselves and IBM: “Why should IBM PureSystems capture my interest?” Then they usually ask where to position PureSystems in their current IT landscape and if the integration will be easy. Let’s discuss these three points.
I have been engaged in many meetings with clients, IT managers and architects during the last six months and the conversation quickly moves to the bigger picture to discuss how their IT should and could evolve.
The fact is that people in business units are no longer passively using the services offered by the IT team. In private life, smartphones and tablets allow intuitive access to a never ending list of app stores and social media so employees are naturally questioning why they don’t have the same user experience and agility when there are back at the office.
I know of a large enterprise with more than 12,000 applications resulting from the pressure of the business to always get new more productive applications. And in smaller companies, it is not rare to have a couple of requests per week for new applications that business people have found on the web and have already evaluated at home during a rainy Sunday.
So, how can the CIO solve the equation? On one hand, a heavy but fundamental set of legacy applications consuming most of the IT budget and time and on the other hand a pressure to provide employees with a “non-disruptive digital experience” when moving from their private life to their professional life and vice versa.
Given budget constraints, there is no other solution than to shift the budget from “legacy to innovation” and this is exactly what the IBM PureSystems strategy is: reduce operation cost to free time for innovation. As explained in my whiteboarding, PureSystems addresses both the agility challenge that CIOs are facing as well as the operating cost issue.
The PureSystems center catalog offers access to a very large set of applications than can be automatically deployed and operated onto members of the PureSystems family such as PureSystems Applications and PureData.
For those clients facing aging, non flexible infrastructures, PureFlex can provide the infrastructure as a service (IaaS) solution that will help them reduce their operating cost while introducing the latest technologies in the datacenter.
In other words, as said one of my clients: “By buying PureSystems, I buy time for my engineers — time they will spend on new projects pulled by the business.”
Now that we have the main arguments to answer the “why?” question, let’s answer the “where?” Often, IT landscapes can be described as follows in the next figure. Possibly hundreds, if not thousands, of applications are running on different infrastructures with different attributes according to the desired service level agreement (SLA). The most demanding applications from both a performance and availability point of view may be running on mainframes while for example ERP may be running on UNIX and collaboration on x86-Windows. Consistently with application life cycles, environments such as sandbox, development, pre qualification, pre-production and production are required for formal compliance with the IT processes.
The answer to the “where?” question is to identify in this landscape a suitable area to start with IBM PureSystems. To identify this area, let’s look at the technical functional and non-functional requirements (FR and NFR) architects may be looking for. This leads to identifying two areas as indicated below. The blue area has a more relaxed SLA and a key requirement for agility while for the yellow area critical SLA and highest possible robustness are keys.
If we now map the IBM Systems portfolio we get the following approach to positioning Pureflex.
So, as a first approach, an answer to the “where?” question is to position PureFlex to address the sandbox, development and pre-qualification environments for applications running on Windows, Linux and AIX. Some customers will only focus on Windows, others only on x86 with both Linux and Windows; others will add AIX. And very quickly, IT teams will start thinking about the Windows and Linux production environments.
Now a key question still remains: “how?” How should you integrate PureFlex in the existing landscape? What are the best practices for such an integration?
PureSystems and PureFlex in particular have been designed to offer a number of integration points with an existing landscape. Storage virtualization is incorporated in the embedded storage. Another of my whiteboardings explains this is here. This means that existing storage can be virtualized by PureFlex to protect past investments.
Network integration is pretty straight forward as IBM Ethernet switches are fully standard while the storage area network (SAN) switches comes from undisputed leaders in this area.
The Flex System Manager presents Northbound REST API allowing integration with the datacenter management tools for provisioning, monitoring, and the like.
My experience is that once the machine has landed on the IT floor, five days are often enough to get it fully usable by the IT team so they can start playing with this IaaS solution. And with two to three additional weeks, the IaaS can be put in pre-production. With a client we even set a Disaster Recovery mechanism between two distant sites by implementing storage metro mirror between two PureFlex systems running VMware.
Let’s now summarize the answers to the three key questions:
- Why? To reduce time dedicated to operating legacy infrastructures and so allow engineers to look at new projects pulled by the business.
- Where? Windows and Linux sandbox and development environments are certainly a good place to start with a new technology. Extension to AIX development or to Windows production are the next natural steps.
- How? PureFlex offers integration points by design allowing quick integration in the IT landscape. My experience is that three to five weeks from machine landing to production can be a realistic objective.
I hope this blog was useful and please do not hesitate to comment.
Learn more about PureSystems at these events:
- February 5, 2013: Webcast: “Smarter Computing: What’s Next. Ready Now.”
- February 12, 2013: Webcast: “Speed Analytics and Simplify Cloud”