http://info.modius.com/modiusblog/bid/41734/Do-Co-los-MSPs-need-Uni...

Here at Modius, we are seeing an increasing number of requests among Co-locations (Co-los) and Managed Service Providers (MSPs) to help them capture more robust and accurate power measurement data. In one sense, this trend is nothing new because all data centers—whether captive inside an enterprise or an outsourced service provider—need accurate power measurement, typically for improving:

• Capacity optimization
• Energy efficiency
• Uptime assurance

But we find that Co-lo’s and MSP’s have a special need that takes power reporting to the next level: Providing disaggregated energy consumption and power usage data by customer at a very granular level, often by rack or even a group of servers. Typically, they need detailed power metering for each customer, principally for:

• More accurate customer billing
• Detailed status reporting to the customer (in real-time) through a customer portal

Customers are now wanting this information not only to be sure their power bills are accurate, but also to try and determine their available power capacity, usage trends, and accurate data to support reporting on PUE and Carbon management. Or even more of a challenge, they need to unify data across different locations because their customers are spread across several different buildings.

Theoretically, some of this data can been captured from the servers. In fact, with distributed systems management tools, reporting on server energy consumption (at the server level) is relatively commonplace. But this data source is incomplete. What if you want to factor in cooling and other related energy consumption? Or what if you also want environmental reporting for bottom/middle/top for each rack? Now, this is much more challenging …

In general, most Co-Lo’s don’t have access to the server instrumentation data at the chassis level. And in terms of power and cooling, we’ve found that most co-location providers are still struggling to unify a broad range of equipment into a single monitoring fabric and extend the framework across disparate systems and locations.