Recently, Enterprise Executive interviewed John Barnard, CTO of the BMC Software Mainframe Service Management business unit. John has 25 years of experience in IBM mainframe systems software management and development in various capacities. He joined BMC in 1998 as part of the merger with Boole & Babbage, where he was a developer for the CICSPlex Systems Management (CPSM) product. He has a bachelor’s degree in computer science, among other degrees, from Indiana University in Bloomington, IN.
Enterprise Executive: Enterprise organizations today are increasingly challenged by the need to provide continual, uninterrupted business service. What do you consider to be the top challenges?
John Barnard: Business continuity depends on the availability and performance of applications critical to the business. These applications have become increasingly dispersed among hybrid data centers and often across multiple geographic locations. Therefore, IT organizations are challenged in their ability to ensure business continuity of critical applications. In one sense, the need to be able to provide continuous, uninterrupted service hasn’t changed for IT in that there has always been a business continuity requirement for applications and services. What is becoming increasingly challenging now is the growing complexity of applications and the number of geographically dispersed data centers that enterprises are managing. The large technology vendors all deliver technology stacks, but enterprises might deploy more than one of these stacks to avoid vendor lock-in. This further contributes to the complexity and provides additional challenges for IT to ensure business continuity.
IT must also manage large-scale applications that are highly distributed, with application front-ends working off of open, distributed computing platforms, and the back-end database and transaction processing being handled by the mainframe. As these applications become increasingly virtualized and deployed in cloud architectures, there are growing problems in the areas of capacity planning, performance management, and continuous availability that exacerbate the problems of co-location of data and data latency. On the distributed side, virtual applications are being pushed down into cloud deployment to reduce cost. In the mainframe environment, the flexibility and maturity of the mainframe offers opportunities to reduce the cost complexity of systems. Companies are also pursuing Mainframe Cost Optimization, which involves reducing mainframe consumption and increasing efficiency through automation and resource-conserving technologies.
As business continues to globalize and the competition for revenue continues, IT service levels and responsiveness will also be an important part of an uninterrupted business service strategy. Social media is one element that’s driving this trend. Someone in one corner of the world can quickly inform the world at large that your company is having a problem with its systems. This poses risks to the business.
EE: Given those challenges, where should IT focus to ensure the availability of critical business services?
Barnard: There are four business areas where IT should focus: application performance management, enterprise capacity and performance management, enterprise event impact management, and IT automation.
Application performance management is important because businesses run on IT in the form of applications that run the business of the company. These applications have to deliver to the end customer or end internal user as quickly and as reliably as possible so IT has strong management and a good handle on all the applications. Along with this, the tools and methodology need to be in place to provision sufficient capacity to run these applications and continuous performance monitoring to ensure that applications are running at optimum levels. Eventually, of course, all organizations are hit by event and impact management. The key here is to immediately react to an emerging event or impact and to be as near-real-time proactive as possible when these events occur by implementing a sound set of predictive analytics on application and system performance that allow you to see trends for workload and application usage around the world. These analytics allow IT to predict what the level of service is likely to be around the world in the future—and to plan accordingly. If you wrap all this together and then proceed with system automation of commonly occurring problem isolation and resolution in a repeatable and standard way, you can reduce the workload and stress levels of your staff and preempt certain problem scenarios from occurring in the first place. An example in a virtualized environment is on the distributed computing side. Here, at any given time, the equipment of the business might not be able to be met by existing resources if they aren’t quite capable of handling the workload—or if analytics is saying they won’t be in the near future. Automation can be used in this situation to deploy the needed resources as long as it’s repeatable, standardized, and scalable—and is able to meet the psychological security need of IT that the process will be done correctly. Many company IT departments have built up their own automation capabilities, but it isn’t clear whether these in-house-developed solutions can meet emerging deployment requirements, such as the need to reduce energy usage, unless they’re broadened in scope.
EE: How do these four business areas ensure business applications remain up and running?
Barnard: One of the ways that application performance management, enterprise capacity and performance management, enterprise event impact management, and IT automation help IT is they ensure that best-of-breed capabilities in all their respective areas are in place for problem detection and resolution—and for the maintenance of high service levels. They also provide tooling around problem remediation that gives IT improved visibility of problem scenarios and the ability to work through problems with mobile devices. The automation that the tools offer can automate workload and resource allocation based upon internal business rules and best practices, and there are also predictive analytics that can foretell future events and problem scenarios. Predictive analytics is still in its early adoptive phase in IT, but analytics have been available for some time in the mainframe environment and are just now starting to appear on the distributed computing side of IT. When these analytics link system alerts to the appropriate set of IT KPIs [Key Performance Indicators], the cost of IT can be driven down and IT also has a tool for gauging its success against the internal business metrics and SLAs [Service Level Agreements] that it’s working to. This paves the way for IT to more fully partner with the end business in the attainment of corporate objectives.