When you are placed in the unenviable position of having to make decisions on selecting a hardware solution for your company's network infrastructure . . . . LAN or WAN . . . . the evaluation process can be overwhelming. Without a plan you're doomed to failure . . . . and a huge migraine.
Keep in mind that it's important to consider the culture of your enterprise and what qualities it values.
For instance, if it values self-reliance in IT - or views it as core to its business - it may be more likely to look for the best of breed solution. If IT is less central to the enterprise, than a widely implemented adequate solution that is easy to find experienced people to work on may be the solution of choice.
In short, making the list of qualities we all want in a piece of gear/vendor is easy. Figuring out which ones to emphasize in the analysis is the real challenge and the analysis that really should drive your decision.
To help things go smoother . . . . . focus on these simple attributes in your evaluation.
The number one answer is the integrity of the corporation. What is very important is that the company has a commitment to make their products work as advertised and documented, and not cut corners on Quality Assurance.
Another important factor, but related , is their customer service and technical support. What is the turn around time to get replacement components that are defective. When you talk to technical support, are they knowledgeable or certified on the product they support and the platforms the product runs on.
Base your provider evaluation on the above and the rest will follow along.
Next . . . . confirm the design requirements, how much network traffic, kind of traffic (data, VoIP, etc. ), number of remote networks for WAN, future growth, redundancy. Cisco has an online tool that will suggest the appropriate device based on answers to these kinds of questions.
If cost if no object you'll do well with Cisco. It may be worth evaluating Juniper and Foundry depending on your needs, and for SMB solutions you may even consider open-source options such as the Vyatta router/firewall.
Over the years I've seen people tackle this question, all in a variety of ways. What I have seen more often than not is the desire to create more documentation / analysis of products / due-diligence without focusing on what's at stake.
Don't complicate the question too much - focus on your specific needs, and make sure you don't exclude the future. What you need now may be just the ice-breaker for what your needs are in the future - make sure you have a plan to scale.
The other big question I think is also overlooked is residual costs associated with purchased equipment. A lot of companies are gung ho on maintenance purchased annually . . . . remember, there is a cost associated with downtime, and in some environments this cost prohibitive; in some it is a non-impact. Factor these things in your evaluation as well as the cost to support the solution.
For a quick checklist:
* First of all evaluate known and proven brands when possible since the issue of continued support from the company and availability of warranty repair and replacement would be a major concern on a significant investment.
* Second - Choose the correct level of product for the job. Avoid paying for added functionality if the client would never, ever (be careful, things can change) use these things. Don't buy a limo when you only need a bicycle.
* Third - Compare performance, price, and mean time between failures (MTBF). Look for “end of life" announcements. If you are looking for a bargain or want longevity these are a good clue.
* Fourth - Google the product(s) in question to find reviews and other feedback.
* Fifth - Hands-on evaluation with a call to support for the finalist products.
Somewhere in here you may need to consider the need for failover or redundancy. If this unit represents a single point of failure without backup . . . . then cross ship warranty policies or local availability may be critical.
Generally . . . . . it all starts with knowing your needs. Routers have the ability to connect networks with different media, even different networking techniques. Examples are Fiber-to-UTP and Ethernet-to-ADSL. It's obvious you should have a device that can address your needs. Will your needs change in the future and, if so, is the device capable of adapting to those changes?
Other considerations are security . . . . . does the device stand at the edge of your network, at the risk of being attacked; or is it somewhere in the middle of your LAN, just connecting departments to the core. In the first case you need something with a firewall feature set, in the second case a layer 3 switch might do.
Don't forget . . . . . what is the amount of traffic the router needs to process.
Once you know what you need, and bring back your white list to the devices that address your needs, more choices have to be made.
When it comes to IT in general, money is a BIG issue. As IT usually will be seen as something that costs money. So at first thoughts, the price of the equipment is important.
BUT. . .
You should consider that also for managing the network environment. When your initial expenses are low but you spend a great deal of time keeping it up and running, it is difficult to adapt to changes, or your company suffers network outages . . . . . your management will not be pleased. So you need to look at MTBF figures, mean time between failure, and how fast you can get a replacement. With some exotic brands replacement can be an issue.
For real important routers you should consider a hot standby configuration which costs more, but will switch over automatically in case of a failure without anyone knowing your primary router died. Except for you, of course, as you are monitoring both devices.
Another important item related to managing the equipment is how it fits in your IT department. If your network engineering department is a group of well-trained Juniper specialists, buying a Cisco brings additional costs for training.
Boiling it all down here's the real message:
Firstly as with all business considerations you must consider the costs there is no point at looking at the top of the market if the business will not stretch to that point. It is also worth discounting cheaper options ASAP if the business is prepared to pay for the right solution rather than the cheapest.
The next consideration depends on the nature of your business, your need for security and reliability. But at a general level most businesses need something reliable. This means if you are remote or have remote offices with little support you want something with a high time between failures. Security often depends on the nature of your business protection. Financial and Medical information is for example considered more of a risk than most general data. There is also always a basic need for security . . . . but again as always there must be a balance of Cost, Usability and Security. It must never be your only consideration. It also depends on the size of your IT support organisation. Will hundreds of people require access to this equipment . . . . or will this be restricted to a select few? Is centralizing and auditing access worth it for your organisation?
Supportability is also part of this equation; you may want something with either great remote management capabilities or something simple anyone can maintain. If you purchase rarer equipment it may be harder to find remote service personnel capable of support. However if you design the systems well . . . . with spares and redundant paths . . . . a centralized body may handle this for you. You see it all depends on your approach to the problem.
Next how high will you scale, do you have growth projections for the future. Are there any new applications or new company acquisitions which will seriously affect the solution. Will you be moving say from a DS3 bandwidth backbone to an OC3 bandwidth backbone in a few years?
Once all the considerations have been looked at you must be consistent. Classify differing sites and have set standards in operating systems, hardware platforms, IP Addressing and configurations for sites. This is great from a TCO perspective and will make supporting the network easier and cheaper. Even when using low end equipment replacing a standard item held it stock is much easier the trying to figure out a new configuration in the heat of a network outage. This also makes documentation easier which is the core of world class architecture. Support on sites without documentation is always a nightmare.
For network designers there are obviously many brand considerations, but most will often recommend CISCO solutions. I can recommend CISCO from a security, manageability, scalability and supportability perspective. However it can be quite expensive depending on your requirements.
It is up to you to manage the balance between price and the rest. You may end up with a different vendor for routing, switching, wireless, VOIP etc. The important thing is to try and keep it manageable. The item price is not the full cost consider Maintenance, Support and Reliability in your equation. Sometimes the most expensive option has a much better support cost than the upfront cheap options.
Whatever solution you choose in the end . . . . hopefully you follow a well thought plan in the process incorporating the above issues and suggestions.
Michael is the owner of FreedomFire Communications. . . .including DS3-Bandwidth.com and Business-VoIP-Solution.com . Michael also authors Broadband Nation where you're always welcome to drop in and catch up on the latest BroadBand news, tips, insights, and ramblings for the masses.