For vendor management platforms to be effective, they need to replace processes IT developed to get around shortcomings of previous platforms.
The management platform must provide all the tools needed for management at a low cost. The management platform isn’t a value add—it’s an integral part of the network buy.
I was talking to a friend who is neck deep in network management in a very large enterprise about some of the new technologies and features network equipment vendors are putting into their network management systems (NMS) with the lofty goal of providing a single pane of glass that has contextual views and workflows built-in. His response, after making a face like he just took a swig of sour milk, was “I’ve heard this all before and each time the platform was expensive and lacked the necessary features we needed. We ended up augmenting with other products. It’s not pretty, but it works.”
One part of the problem my friend faced was that network management products were long on promises and short on delivery. Device discovery was never 100% and network mapping sometimes resulted in impossible, Klein bottle style topologies which had to be corrected by hand. Another part of the problem was self-inflicted. His IT staff would use the CLI or custom built scripts to manage network elements which resulted in the NMS being out of date and constantly needing updating. A whole cottage industry developed to rationalize the NMS view of the network with reality but that’s because the NMS’s weren’t effective tools for IT. I think that’s changing. Continue reading “The Changing Face of Network Management”→
Intelligent embedded network agents and sophisticated software heuristics provide key insights into information and performance patterns for predictable data consumption, but interpreting these requires talent
Humans remain the most valuable troubleshooting tool in the IT arsenal
Having worked in infrastructure in the ‘90s and I’ve done my fair share of troubleshooting vampire taps, thick-LAN, and eventually thin LAN (and those finicky terminations) I can say we’ve come a long way. Granted at its most basic we’re troubleshooting low voltage electrical wires in most wired infrastructure. Sophisticated tools are embedded in many switching platforms now which can immediately detect a link loss in addition to whether it’s a damaged cable or connector, or alert correlation from multiple devices to pinpoint the exact location of a ‘noisy’ device polluting the network. Advances such as these have increased efficiency, reduced trouble ticket resolution times, and freed up valuable resources to work on more complex challenges. With wireless access becoming the norm for clients as more and more devices go solely mobile, tools have generally kept pace and network management systems have slowly grown more capable and feature rich. As cloud adoption rates increase and systems grow more diverse, the tools are likely to suffer a setback, though, with many disparate elements, both physical and virtual, contributing to a single application connection. Troubleshooting these will once again require a significant amount of technician involvement to determine root cause during an outage (and no, rebooting your client isn’t the answer, Mr. Helpdesk). Physical and virtual agents must be deployed in order to collect statistics in real time and aggregate these bits into a collective perspective of the health of the network. Whether this is done with one of the extensible “framework” NMS systems or via vendor element management systems does not matter, but at the heart of this is that enterprises need to embrace a more sophisticated management model than they have in the past. Continue reading “IT Pains Evolving: Where’s Holmes & Watson?”→