Complexity in data centers

IT infrastructure has transformed itself from being a necessary evil to that of a key business enabler, helping companies develop solutions to differentiate them from their competitors. IT infrastructure in modern day enterprises is the backbone that helps them stand straight with their head above the competition. To this effect, the data center landscape, which hosts this infrastructure, has evolved from a few servers in an obscure corner room of a building to that of thousands of servers in different buildings spread across various geographies. The technologies deployed in these data centers also have transformed from Mainframe and Unix systems, running e-mail and legacy applications to heterogeneous, distributed solutions involving database, middleware servers, Commercial off the Shelf (COTS), packaged, and custom applications. Further, these products and solutions interact among themselves to provide external facing business services and enable day-to-day internal business operations. The advent of Web 2.0 and cloud computing and niche features such as Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) have further complicated the landscape.

The following image shows a functional view of a typical enterprise IT infrastructure:

The infrastructure consists of both external and internal applications serving various classes of users. These users access various applications through different access points and devices. Even though actual IT infrastructures are far more intricate depending on the business domain of the enterprise, the above minimalist view clearly demonstrates the complexities involved. To this view, if we add the collaborations among the various entities, the topology becomes almost unmanageable. The following is a very simplistic illustration of the physical topology of the infrastructure that supports the earlier functional view:

It can be seen how, IT impacts every aspect of the business operations—ranging from customer care to end user interactions to accounting to employee self service. Needless to say, the performance of the IT infrastructure is a key driver towards the success of the enterprise business.

This complexity in the IT landscape necessitates deployment of a highly sophisticated management solution across the enterprise. Such a solution must be able to manage all aspects of the IT infrastructure, starting from physical hosts and devices to packaged applications. While the solution should definitely cater to managing disparate components individually, it must also provide visibility into the complex business processes and usage of the underlying infrastructure. The former view is required as a tool for day-to-day IT operations by system administrators and support personnel who know the physical topology very well. The latter view provides the CXO-level senior management with invaluable insight into the effectiveness of the underlying infrastructure in driving business operations.

Many of the applications and business processes interact with each other and come together, to provide meaningful services to both external and internal users. Such interactions are achieved using diverse technologies and architectures such as SOA, web services, cloud computing, Web 2.0, and so on. These services must also cater to the availability and performance expectations of customers and internal users. These expectations are formally referred to as service-levels. The commitment on availability and performance of these services, commonly referred to as business services by the service provider, is defined formally using Service-Level Agreements or SLAs. Enterprise-wide management of these business services including their service levels requires technology-independent perspectives that provide the CXOs with the big picture. The above management concepts fall under the broad category of BSM.