Achieving an enterprisewide view requires overcoming significant problems at the technical level. There are more complexities in the distributed systems environment due to the mix of virtualization, the greater number of systems, the interconnections, and so on. Software as a Service (SaaS), or cloud computing, is primarily oriented to the distributed side. In addition, information technology semantics vary, depending upon whether the language of discourse is for distributed or mainframe systems. It’s important to address these questions:
- How does SPECint compare to MIPS?
- How do you measure hypervisor overhead compared to system overhead in z/OS?
- What’s the issue with network delay on the mainframe and why do distributed teams make such a big deal about it?
So, there’s a problem with semantics that occurs at the monitoring and measurement levels, as well as with trying to understand what the applications and characteristics are as information crosses the boundary between the distributed and mainframe worlds. In the mainframe world, the boundaries may be defined as a z/OS box, for example, running an operating system with multiple instances of subsystems and middleware layers. The challenge involves bridging that gap the distributed and mainframe sides.
Communication is critical. Mainframe systems management and distributed systems management staff often work in silos. They need to communicate to understand the problems each has. After all, the applications cross all the software silos. That’s really an organizational problem, not a technical one. Addressing this cultural issue requires realizing that the business is agnostic in regard to distributed or mainframe systems. Business applications and business processes are agnostic, too. They go everywhere in an enterprise; they participate equally in the business’ success or failure.
Aware of these realities, information technology organizations are beginning to view their infrastructure— both distributed and mainframe—as a single entity. However, the organization needs to catch up with this concept. While this type of recognition is difficult to achieve from an organizational perspective, it can be quite simple— with the right tools—from a logical, technical perspective.
One of the prerequisites of viewing the enterprise as a single entity is that tools and tooling can also be considered somewhat agnostic, at least to the extent they support both distributed and mainframe systems. Often, organizations have a set of tools that work on the distributed side and therefore are used for distributed monitoring, system management, or change management. On the mainframe side, tools provide the same management functions, but behave differently partly because of the different semantics. Each side requires a different type of subject matter expertise. Many organizations strongly value shared tools. Applications that span mainframe and distributed platforms are extremely important. Vendor management solutions that provide common semantics across all aspects of the information technology enterprise are invaluable.
Understand Before Taking The Next Step
In a team sport, players need to understand the other players’ intentions. For example, when a catcher signals a specific pitch to the pitcher, the pitcher needs to understand the signal and either acknowledge it or indicate disagreement so the catcher can signal a different pitch. The same is true with communication in the distributed and mainframe world.
It’s important for IT personnel to understand the relationship between the mainframe and distributed systems as it relates to business applications. This is where semantics are important. IT personnel should understand how data on both the distributed and mainframe sides gets converted to information. They should ask, “How can that same information be represented on the mainframe side? How can that same data represented on the mainframe side— with different data types and metrics—be used to generate information that semantically is the same as what’s on the distributed side?” These questions help identify how IT personnel can ensure that the different types of data on the mainframe and distributed sides come together to form the information that’s useful for both information technology organizations and business. Even though the data may appear different on the distributed and mainframe sides— given the nature of application flows—these are simply different ways of representing the same information.
At the data level, it’s important to understand the demarcation line between the distributed and mainframe worlds. In the classic mainframe world of the z/OS operating system and its subsystems and middleware, the data boundary may cause a disconnect at the technical level between the distributed and mainframe process. That gap has proved to be fairly difficult to cross in terms of providing an integration between distributed and mainframe processes.
How do you solve the technical challenges related to semantics and then understand the demarcation line between these two environments? Solving these challenges requires creating common information out of varied data. Business Service Management (BSM), a comprehensive approach and uniform platform for running the IT department, is an effective way to accomplish that objective. With BSM, information technology organizations have a comprehensive view of their entire enterprise, including systems, database, middleware, and transaction management, as well as workload automation, batch optimization, and other areas.
Solutions on the mainframe side should provide alerts or alarms that could be used to reflect service impact on the enterprise. For example, solutions can flow that data to a Configuration Management Database (CMDB), where it can then be used to generate alerts at the business service level. That process should be in place to take uncommon data and create common information from it. This approach helps to address part of the semantics problem.
Another aspect of the semantics problem involves addressing the difference between business applications and technical applications. At the business level, the data underlying the applications flows can be surfaced with sufficient tooling to present business information—whether it’s number of orders, dollars, policies, or other metrics. That information can be associated with the technical metrics and data to understand the performance and availability of all business applications.
Play To Win
Viewing the business environment in the context of the whole enterprise can reduce costs, leverage skills, and save time. A unified solution for monitoring middleware and associated business transactions across the enterprise can offer the depth and breadth of coverage needed to deliver improved service at a lower cost. By converting diverse data to common information, you can facilitate the teamwork between mainframe and distributed systems IT personnel, and IT department and business personnel. This teamwork can help your organization meet its current and future challenges. ME