* Closer look at Statscout’s Performance Monitor
Statscout’s Performance Monitor customers use close to 100% of its functionality. This may not sound like much, but as far as I know, it’s unique – since Performance Monitor addresses a range of strategic requirements to monitor networkwide performance, including the performance of large and often complex networks.
With most strategic management products, 20% of feature utilization is considered a successful deployment.
Statscout’s Performance Monitor was built on the principle that, given the overwhelming nature of infrastructure alerts, less – if it’s well chosen – can be more. By limiting the collection of SNMP Object Identifiers to highlight critical statistical values, Statscout minimizes the traffic over the network so that the burden on the network is minimal.
This enables what Statscout calls “blanket monitoring,” so that virtually every network device and every port in the network can be monitored. There’s no guessing whether a networked device is critical or not. There are no blind areas in terms of devices and ports, period. In one example, a customer network with 288 devices and more than 30,000 ports with frequent, 60-second sampling experienced less than 1M bit/sec of bandwidth utilization for network traffic. Statscout historically has extended to environments with as many as 200,000 monitored nodes, and Statscout claims it can scale much larger.
This efficiency in data gathering enables another fundamental design feature within Statscout Performance Monitor. Statscout’s analytics can function at full data utilization to prioritize relevant data for visualization. Statscout uses real-time data sorting to present the most likely issues at the top of a view. As a rule, even in very large or complex networks, there are generally only a small number of links or nodes that show irregular statistical behavior at any given time. Also, the typical behavioral characteristics of potential problem areas are usually conspicuous in contrast with other elements or networked areas that are operating well. This enables a simple sorting of prioritized information even in large environments.
It should also be noted that this design point is built around normalcy and deviations from normalcy, as compared to fixed thresholds. This is another critical Statscout approach – a deep mistrust of thresholding in its traditional sense, when high thresholds often filter out critical alerts and low thresholds often flood the administrator with so many alerts that the data becomes all but useless.
Enterprise Management Associates is seeing a number of technologies and innovations emerging in the market to address this issue as a more “intelligent” way of managing events.
Statscout’s historical reports are available as an archived version of its real-time reports and provide added context for real-time analysis and troubleshooting, as well as for network planning in general. The focused data collection also limits the data so that historical records can be maintained without loss.
Statscout customers have validated another feature: ease of use and administration. Performance Monitor, while it requires some time to establish a baseline, works well out of the box, collecting useful information within minutes of booting up the system.
The combination of blanket monitoring across network devices and ports, focused data gathering at frequent (60 second) intervals, automated baselining and overall simplicity and clarity are the lynchpins of Statscout Performance Monitor. It supports hubs/repeaters, bridges, switches and routers and servers with Windows NT, Unix and NetWare.
By design, it does not provide full-service infrastructure management, and is focused on the performance of the networked infrastructure and its interconnections – as opposed to, say, detailed analysis of applications and application dependencies, or databases. It is also, by design, relatively fixed in function. But it works well, and at a comparatively low price. Initial installations of Statscout Performance Monitor for a 1,000-interface license are at $15,000, with many full network installations running about $35,000. As such, SPM provides an excellent, usable and accurate set of insights into network performance.
