IT management efforts to focus on service delivery within the enterprise are hamstrung by traditional IT management tools that view the world from a bottom-up perspective — a method of managing individual components within a network under a specific category or silo. What’s needed is a top-down perspective.
While element management systems do a fair job monitoring assets in a specific silo, there is no way to manage the interactions among endpoints (both client and server), applications and networks across silos. No matter how you fine-tune these silos, it is not enough to combat the day-to-day availability and performance problems that inevitably crop up.
This has caused a plethora of challenges, from failure to identify lingering issues behind application performance problems to difficulty pinpointing the source of specific problems identified by users. It is often difficult for IT to find problems because the source is obscured in a complex web of interactions.
This is particularly true when the help desk escalates a trouble ticket to a silo without gathering additional information. An IT engineer using specialized analysis tools might then conclude there is nothing wrong in that particular environment, and pass the trouble ticket on to another silo.
This escalation model can be time consuming given the widely heterogeneous nature of enterprise networks and management systems. It is also likely that this scenario will only get worse as individual silos roll out new applications and endpoints — often without the knowledge of the other teams. With the demands on enterprise networks evolving rapidly, a radical change is needed in the way infrastructure is managed across the company.
A top-down approach is the only way to revitalize the role of IT management. This approach views IT assets as they are in the network — a set of interwoven, interdependent endpoints that produce and consume applications and services that run from anywhere, at anytime across a common infrastructure.
With the top-down management concept, endpoint devices and their applications are at the top, providing a clear picture of enterprise wide performance issues by understanding the endpoints and how they interact with each other through applications.
Top-down management begins by gathering details about endpoints and applications, how they normally behave over the infrastructure and when they are adversely affected. IT management can bridge silos to correctly reflect how IT assets have been deployed and are used by the organization. Every infrastructure object has the potential to affect not only the business but also other objects, not just in its own domain, but in other domains as well. Application performance is managed in conjunction with the other parts of the infrastructure.
To understand how a top-down approach would work, let’s examine a common enterprise issue — a dropped call in a VoIP environment. There can be a variety of reasons why a VoIP phone doesn’t function as expected. It could be the device, but more likely the problem stems from the overall VoIP ecosystem.
Unlike a bottom-up approach that would first examine each silo (such as the CPE and the voice mail application), a top-down approach uses flow information to automatically discover the IP endpoints involved (such as the IP phone, IP call manager and IP voice mail system) and the applications involved (such as call signalling, the RTP voice stream and voice mail) on the LAN and WAN.
As part of this discovery process, each endpoint is associated with a specific location (topology), and all application usage is examined across endpoints.
That makes it possible to develop a precision profile for each endpoint, application and location across a range of variables: such as bit rate, packet rate, burstiness and affinities. This gives IT an established profile of each endpoint, location and application, so when a problem occurs, the “established” profile can be compared with the application operating in real time to determine the source of the deviation.
The solution then continuously compares each profile with the real-time application behaviour of the element to determine how the interactions in and surrounding the ecosystem changed when the problem — the dropped VoIP call dropped — developed.
The result is a real-time detection of any atypical endpoint or application behaviour, providing a view of these symptoms across time, application and location to identify the core of the VoIP problem.
Continued dependence on a bottom-up approach to IT operations will cost enterprises millions of dollars in lost revenue and productivity annually. The Yankee Group estimates that workers lose approximately 35 days of productivity per employee per year due to poor performance and availability of applications.
The top-down approach of infrastructure management leads to a model in which IT is the first to respond to problems, not end users. Once all the pieces are in place, IT groups can move toward business productivity management as opposed to a reactive element management.
Messina is vice president of marketing at Xangati.