Infrastructure Management vs. Service-Level Management
As enterprises have become increasingly dependent on services provided by software applications to function, managing the quality of these services in terms of parameters such as up-time, performance, turnaround time, security, and consistency has become ever more important. The business criticality of services is such that any disruption is no longer treated as an IT problem; service disruptions are immediately brought to the attention of the CXOs of the enterprise. The CIOs mantra today is to reduce planned downtime for the business critical services to zero—unplanned downtime or service disruption is a cardinal sin.
IT management today has been forced to rethink their methods of managing and operating IT infrastructure. For a long period of time, IT managers delivered against infrastructure-level SLAs, such as “if server XYZ had 99.9% availability” or “if network segment DEF had an average latency of less than 30 ms,” all was well. The standards for management, the operational processes, and the expertise built within IT management teams was geared toward the management of individual components of infrastructure.
However, IT management teams today are increasingly having to deal with SLAs at the level of application services and sometimes even business services. Today’s SLA statement is along the lines of “the hospital admission workflow process should have less than 1 hour of downtime a month” or “the average time taken to automatically adjudicate an insurance claim of a specific type from a specific provider should be no more than 20 seconds.” There is an enormous difference between the availability of an admissions process and the availability of a server. It is the difference between having to manage application systems and managing a piece of infrastructure that in the context of an application system is merely one of several parts. IT management teams today are struggling to use the existing techniques of managing infrastructure to cope with this difference.
Complexity of Managing Application Ecosystems
Twenty years ago, the applications that enabled business processes ran on mainframes or on client server systems. From an IT manager’s perspective, the number of moving parts were small and the interconnections between individual applications were limited and most often were based on one-to-one data exchange mechanisms. Management of these systems could be performed on an individual basis. Recently, this scenario has changed significantly.
Today’s application systems are fundamentally different and significantly more complex. On one hand, it is fairly common for applications of today to have thousands of objects. On the other hand, applications seldom function in silos; instead, they are integrated using a variety of application and process integration mechanisms. Application integration architecture focuses on the creation of enterprise-level platforms or integrated fabrics that subsume individual applications and the infrastructure they run on. Instead of individual applications, one has an integrated platform that exposes a set of “business level services” used to compose business process workflows. This translates into a large system that has many moving parts and a complex web of dependencies between these parts. It is no longer feasible to manage individual applications without considering the connections between applications.
These integrated platforms operate within environments that greatly add to this complexity. For example, the environment almost always has a set of multiple standards such as BASEL-II or ITIL that the business processes and the applications and the teams that manage them must adhere to. There is a large variety of direct users who interact with the applications; these include stakeholders, such as customers or suppliers, and business managers—any outage or degradation of performance directly impacts them and, hence, the business. Outsourcing business operations and IT systems management, global accessibility requirements, virtualization, and the need to scale on demand contribute their share to this complexity. The word “ecosystem,” which brings to mind organic complexity and the notion of several entities coexisting and interacting with one another, is singularly appropriate to describe the application system environments of today.
Application ecosystem: The complex web of closely integrated software applications, the infrastructure they operate on, and the processes and standards that govern their design and operation. Application ecosystems provide a set of “business level services” whose quality must be managed.
Comments are closed.