Solutions
The Challenge – A Gaping Hole in Network Visibility and Awareness
Problem 1: I don’t know how my services are flowing through the network.
One of the key challenges of supporting critical application services is the lack of visibility into how these services flow through the IP network and relate to specific network elements (e.g., routers, links). The complexity arises from the way that control over network behavior is distributed across many network elements (routers) that all work independently and react in real time to changes in the network. While creating resilient networks that automatically recover from failures, this distributed control makes it very difficult to trace how a given service flows through the network at any point in time. As a result, the process of isolating the root causes of service disruptions and performance degradation that originate in the network becomes incredibly complex, costly and time-consuming.
Protocols have evolved to provide greater control over how traffic flows through the network, but those controls have introduced additional challenges for IT staff, such as the significant distortion of traffic patterns resulting from simple configuration errors. Consequently, the operational costs of maintaining end-to-end IP service quality have increased dramatically.
Lack of visibility creates another challenge: the integration of change and configuration management across both the network and the applications/services it supports. Many enterprises have instituted formal change management process for their critical applications to ensure that they are not disrupted by untested changes. However, without visibility into how services flow though the network, IT staff cannot properly apply these policies to the underlying network infrastructure without labor intensive, error prone and rapidly outdated manual service mapping. As a result, these network changes often cause unexpected service disruptions – accounting for as much as 30% of service problems.
Problem 2: The quality and performance of my applications are degraded but I don’t see anything wrong with the network.
To date, most network management tools have focused on persistent network problems because automated network recovery from local failures was sufficient to support traditional applications. These tools relied on periodic polling (e.g., every 5 minutes) to check the health and status of individual network elements as well as end-to-end service quality (e.g., latency, jitter, packet loss).
The problem with this approach is that it is blind to short duration network disruptions that can trigger changes in network behavior and wreak havoc with time-sensitive applications. As a result, the network can be experiencing problems that severely degrade or disrupt critical service delivery without anything being detected by traditional network management tools. With no visibility into these problems, it is impossible to isolate and resolve them in order to drive continuous improvements in service quality. This is significant because, according to IDC, 50% of all application performance problems are caused by routing instabilities and anomalies.
Next: The Solution