Guest Post: Searching for the Mythical Single Pane of Glass

glass dome By mendhak

glass dome By mendhak Written by Floyd Strimling, cloud technical evangelist at Zenoss.

IT specialization promised efficiency and increased performance. A couple years ago, it was common practice to keep the applications, server, storage and network teams separate from each other. In fact, IT managers purchased categorically distinct software packages because collaboration between these teams was not normal. Today that legacy still holds true, albeit with more silos and more complexity. It’s not uncommon to see separate silos for compute, virtualization, environmental and cloud teams in addition to the four traditional silos mentioned earlier. Complexity begets complexity leading to a nightmarish collection of separate IT departments with redundant services and a tangled web of systems that are hard to decipher when things go wrong.

In the beginning, Elemental Monitoring Solutions (EMS) came from vendors themselves and were just that – elementary. The management capabilities in their tools were always an after thought and failed to deliver the usability needed to make sense of events. EMS systems weren’t the best way to manage systems, giving rise to third party niche solutions, which were merely extensions or augmentations to the EMS solutions.

To fill the need for encompassing monitoring solutions, big vendors such as IBM, CA, BMC and HP started providing frameworks that promised everything would work wonderfully, but these proved to be expensive and they lagged behind current versions of IT infrastructure they looked to monitor. Into this vacuum of affordable and capable monitoring tools came homegrown solutions that were risky and not always in line with IT standards.

While IT departments tried to figure out the best solutions to monitor and manage their infrastructure, the business had to keep on rolling, leading to “frankenmonitor” solutions, those hodgepodge collections of tools slapped on together by different IT groups for their own silos.

Event Management tools were supposed to solve this issue, but they were purpose built for static network infrastructure. With these tools, events from disparate systems are all funneled through gateways and interconnections and add correlation information to pop out root causes and tell operations departments what’s going on. Unfortunately, these proved to be limited in their capabilities because event management technology can’t be fast or rich enough to handle today’s dynamic infrastructure and workloads.

The follow-up to event management was Configuration Management Database tools (CMDB), but these pulled information from separate databases that often had stale data. Again, this type of technology proved to be too slow to keep up with the rapid rate of change found in most enterprise IT departments. The Holy Grail is a single pane of glass that provides IT managers with a single access point to make sense of their entire underlying infrastructure in real time.

What today’s operations teams need is a simple platform that monitors events across infrastructures and provides a new way to analyze and plan for impact management. There are 3 important consideration to take into account when chasing that elusive single pane of glass; 1. A solution should monitor all kinds of devices with zero touch and no agents to deploy 2. These solutions should be open and malleable so developers can extend its capabilities to do exactly what they want it to do. 3. A solution needs to be a unified product that is not simply wrapped and rolled with other little known solutions. This is important if we’re to break down silos that can holistically tackle an enterprise’s infrastructure.

We need to move beyond gauges, charts and pretty screens within separate silos and start looking at systems holistically from a services perspective. A services perspective means looking at all the services that are running to gauge health instantly. This requires interconnections between disparate pieces, what a CMDB is supposed to do, but in real time and at a deeper level.

Floyd Strimling is a cloud technical evangelist at Zenoss. Floyd has been following the cloud computing/autonomic computing (and predecessors), datacenter automation, virtualization, networking and security areas now for over a decade. Floyd also writes on technology trends at his personal blog, The Platen Report.

Photo by mendhak