Seemingly every day another vendor launches a service that promises to revolutionize the way they monitor their IT infrastructure. Generally these launches comply with all the buzzword – cloud, big data, predictive analytics etc. While I’ve no doubt that IT infrastructure monitoring is vitally important, it seems that vendors, in an effort to gain extra attention for what is essentially a boring service, are jumping on the bandwagon and using as many current words du jour as they possibly can.
This week’s contribution comes from Zyrion, a Cloud and IT monitoring vendor who has released their predictive analytics capabilities for the monitoring of distributed clouds and infrastructure. The core premise is sound – more and more organizations are moving to a broad heterogeneous IT infrastructure setup – in doing so they want to get monitoring visibility across all of their kit. Add to that the promise that these solutions will use machine learning to identify patterns and you have a proposition that promises the holy grail – continuous improvement and ever increasing fine identification of performance degradation. The premise goes that IT infrastructure, once fairly static in nature, is now becoming ever-more dynamic – monitoring platforms need to keep up with this shift.
This launch follows last year’s launch of the Zyrion ‘Data Capture and Processing’ module for seamless monitoring of Cloud technologies,and the 2011 release of the ‘Automation’ module. This third phase of Zyrion’s self-titled “platform evolution” aims to deliver features aimed at reducing false alarms using predictive analytics in IT environments.
Specific features in Zyrion’s predictive analytics module includes:
- Behavioral Pattern Analysis: Automatically determine the behavior of any and all IT components by analyzing the performance metrics for time of day and day of week behavior over any given period of time
- Flexible Baselining based on Behavioral Pattern: Administrators have the flexibility to adjust the calculated baseline behavior using any statistical calculations of mean, peak or 95th percentiles
- Composite Thresholds: Allows creating composite Service Metrics for any IT service and modeling the behavior of this Composite Service Container metric
The proposition is sound – especially when one looks beyond the buzzwords! Heterogeneous IT monitoring is one of the big overarching themes I see coming to prominence over the next few years – while there’s a lot of players in the space, and some will obviously fall by the wayside, those who knuckle down and deliver on the promise look set to capture a big opportunity.