Posted by Chuck Bessant
Air travel has come a long ways since Wilber and Orville gave it a go at Kitty Hawk, NC. As air travel matured, planes became more sophisticated with the assistance of avionics, or electronics designed for aviation. Humans are still required (even for pilotless planes, someone is controlling them from a remote center) to fly aircraft. The more that is expected out of an aircraft’s mission, the more sophisticated the avionics. In addition, avionics allows for less flight crew members resulting in lower labor costs. Imagine trying to safely fly a Boeing 777 with no advanced avionics.
Large enterprise networks are complex. They span hundreds or even thousands of locations over a variety of vendor equipment and operating software, service provider connections, private connections, fiber, copper, microwave, transport a variety of protocols, require different service level requirements per application and/or customer, transport timing sensitive data, must block malware and nefarious attacks, and be reachable by thousands of internal (employees) and potentially millions of external (customers and partners) users. They are depended on for reliability and performance to carry voice, video and data each with unique performance characteristics. As each user conducts a transaction to view email, place a VoIP call, move an MRI to a radiologist in Australia for evaluation or cerf the web to order the latest footwear from Zappos.com, an expectation is set that each transaction, regardless of type, will be completed properly and efficiently. The expectations and reliability of data networks today is that of a POTS (plain old telephone system): you always have dial-tone and the call is always completed.
Since starting back in the early 1980s with networking at Digital Equipment Corp (DEC), I've always viewed management of networks as a critical component to a successful environment. Since the days of thick-wire Ethernet with brick-sized transceivers and X.25 circuits, the complexity and dependency we have placed on networks to conduct day-to-day activities has increased exponentially to the point without proper management, business would collapse under excessive labor costs associated with monitoring, assuming people can perform as well as computers. Service level agreement (SLA) penalties can also kick in (kick you?) when SLAs are breached.
OSS (Operational Support System) was originally a collection of systems architected and designed to support large, complex networks for telcos and ISPs. It is impossible to manage these infrastructures using similar "seat-of-your-pants" technology Orville and Wilber depended on for flight, albeit a short and safe flight.
OSS follows a basic fault, capacity, accounting, performance and security (FCAPS) model. Pretty much everything you need to know about managing a network but were afraid to ask. Riddle me this: Why is it that enterprises look at OSS as "it's a telco thing" and feel they don't need or want it? I've heard this description of OSS many times, and to be truthfully I find it annoying. Excellence should always be sought, even if it is a “telco thing”.
In many ways, enterprise networks have become more complex than many ISPs with the use of WAN optimization, compression, encryption, intrusion detection systems (IDS), VoIP, VPNs, regulatory requirements, HSPDs (Homeland Security Presidential Directives), multicast, caching, malware detection and efficient use of capital equipment to contain cost just to name a few.
Call it network management, Frick-and-Frack or whatever the kewl expression du jour is, it should still cover FCAPS - a simple and elegant model that works and does what people expect. How it is architected, designed and implemented is for another discussion. Each enterprise will have different technical and business requirements, yet each need to cover FCAPS. The magic is in how it is planned, architected, designed, implemented and maintained.
Network and application performance is a direct reflection of the perception of quality the enterprise brings to all users. Without the proper tools, systems and processes in place, service, quality, security, and performance can't be monitored, managed and tuned.
Who's on the network? How close is capacity saturation, or before users slam into a mountain they didn't see on the horizon? Who’s using the resources? Has unauthorized accessed been gained? Have systems been compromised? Outage isolation and problem determination - things break and these need to be found and fixed quickly.
NOCs (Network Operations Centers) are staffed by support engineers that are (should be) more sophisticated than the Vikings in those great Capital "What's in your wallet?" One commercials. To properly do their job, engineers need the right processes, policies and tools, in that order, already in place to help avert a disaster, and to deal with disasters when they occur.
Tools are used to facilitate the processes and policies that were developed based on business requirements. No one tool fits all. Careful selection and integration must be made after close evaluation to ensure capital investment is not wasted. A mix of tools from a variety of vendors will be required.
The sophistication of OSS deployment will vary based on business and technical requirements that are appropriate for the required and offered services. Good analysis will help determine the level of sophistication required to properly and cost-effectively manage services offered by an enterprise. There is no silver bullet or single vendor approach to OSS. It can be complicated yet eloquent and actually save the business money. Less staff can support more, yet not be a burden on the support and design engineers - not too shabby, and shareholders will appreciate it too.
You can call it network management or you can call it operations, but don't call it unnecessary. Try to think about OSS as not just a “telco thing”. Perform a gap assessment using FCAPS as a guiding light, and see if your business can has what it takes to survive growth, an outage or disaster.
OSS most certainly applies to enterprises, and if done properly will enhance the technical, management and cost effectiveness of the business.
Chuck

Comments