Professional Sysadmin Experience • Have deployed dozens of NMS’ over the last 10 years • I work in more of an engineering focused environment than IT focused one. • Don’t do windows, don’t ask
although there is a lot of personal experience involved. • There are very few right and wrong answers on how to get to a well managed network, but I’ll offer a few. • The topic of management has been mostly a shrouded art during the concept of an NMS. • Yes, I know I’m a systems guy and there’s a lot of network talk. Remember McNealy said “The Network is the Computer”. • Let’s talk a bit...
there was Cabletron. • What did a network look like then? • Primarily single-vendor deployments. • No seriously, there’s nothing beyond layer 2. Move on. • Your average corporate network didn’t have a PC at every desk • ZERO Convergence, lots of fat stacks (VINES Anyone?) • Networking focused on how you adapted around the CPE
gobbled up companies that did early management systems or they grew them in-house, such as Cabletron’s Spectrum. • The NMS came from your vendor and it was specialized for your deployment • They largely didn’t exist outside large operations • Commercially available packages were prohibitively expensive, and I hope you don’t want patches with those... • Standards were kept in the realm of ISO and the ITU, go away little guys!
average human being • Trumpet Winsock (1992? 1993?) • Geocities (1994) • ICQ (1996) • BSD, VAX, and this new guy called Linux are all the rage! • Commercial network convergence makes leaps and bounds with Windows NT
concepts of Network Management fit the models of the vendors who sold the networks, interoperability was not of much concern. • Interoperability standards were defined by the ITU and to a lesser extent, ISO. • New standards orgs such as IETF and ICANN operated in a vastly different manner, and therefore, different languages were spoken by different parties. • Standards? Sure, you can have them, for $50,000 to start!
• Author of eight books, including “Exploring the Internet” amongst others. • Created first Internet Radio Station • Responsible for publishing more public data online than any other human could possibly be credited for in their lifetime. See public.resource.org for everything. Literally. • Currently working on digitizing every piece of government video ever, as well as many other projects.
title Tony Rutokowski circa 1990 and continues conversations with this follow through the time we’re talking about now. • All Telecom standards at this point were kept in the ITU’s Blue Book. • Carl realizes the absurdity of unpublished standards. • Carl realizes the ITU has no determinate legal basis for asserting copyright over the 20,000+ pages of the Blue Book. • He makes the ballsiest move in the history of the Internet...
to the point where Carl says he’s going to get a copy of the Blue Book, scan every page, run each one through OCR software and publish it all to an FTP server. • The “threat” in his own words were that someone was making a lot of money from the Blue Book sales as it cost about $1 a page at the time and that it had to impact someone’s bottom line. • Does not backfire! Mr. Malamud finds himself in Geneva, by invitation, a year later to free all the ITU’s documents. The ITU tells a different story than Mr. Malamud’s book “Exploring the Internet”. They word it as an “Experiment”. In 2007 it starts, and continues till this day. • Important for us? The papers defining FCAPS.
conversations in the same language. • We as admins can all focus on how we adapt our practices to meet published accepted practices. • Conversations happen between ourselves and our vendors in this same language. • Little guys can now be involved and play catch up. • Freedom to read the most boring self referential documents ever written!
now have sanctioned interoperability, vendors have no need to maintain management systems in their stacks. The Layer 1 and Layer 2 cartels are dismantled. • Big expensive management systems are still in place, there are no replacements for telecoms... yet. • NMS Platforms are spun off from the parent companies, such as Cabletron’s Spectrum being sold to CA. Many companies end up going defunct completely. • A new era, but there’s this huge void!
the game in infancy • SolarWinds • OpenNMS gets spun off from Oculan • This little app called Netsaint • Neat graphing tool called MRTG to measure “adult content” • COUNTLESS OTHERS
a relative infancy • A veritable explosion of management options • Open Source projects are forked like crazy • Lots of new players • Commercial • Open Source • “Open Core”
logging as possible • Flexible alarming and notification system • Agentless • Reports for management, simple view for NOC. • I don’t want to know it’s there • Scalability and flexibility
with flexible retention. • Modular notification system (It can tweet!) • No arbitrary limits on scalability • Based on open standards like SNMP • Open Source model insures interoperability with my vendor’s equipment and other management apps. • Baked in Jasper reports
login at all. Choose your level of involvement. • Tight coupling of the concept of performance data being linked to faults. • A single go-to place for operational data including • Historical data for faults, performance, and events. • Asset information • SLM by arbitrary groups such as equipment by vendor or by location
• Most importantly, it was flexible enough for everyone to have input on from operations, engineering, NOC staff, and management, and let us all collaborate on issues in simple terms in the most efficient way possible. I needed something for everyone, not just admins.
happy with what you have? What’s wrong with it? What’s good? • How many different products are you using to meet these goals? • What’s the technical and non-technical levels of involvement at your organization around these products?