Could I suggest a path which is easier to say than to do? The first step is an inventory of what you have, the second step is a measure of the QoS you are getting in real terms… and if you can do one and two then three should be to figure out what you really need to improve or eliminate.
I have noticed that throwing bandwidth at problems is a very typical NA approach whereas in Europe bandwidth has always been more expensive so enterprises have tended to be more efficient.
One area you may care to consider is that if you have T1, E2, fractional, DS3, OC3, MPLS etc (oh let’s throw in cellular, VoIP, TDM and any other part of the alphabet soup) …… you probably don’t have a network architecture and you certainly don’t have a central “entity” responsible for optimising it.
So, inventory, contract review, shorten/lengthen contracts to achieve co-termination and use this time to build a real solution and an effective competitive RFI (ask the vendors for their best ideas) an RFP (marry the best ideas of all vendors) and a final contract. Bearing in mind the 80:20 rule of using a “wild card” supplier to keep the main vendor under some degree of competitive control.
The other piece of the puzzle that has to be consider also is the use of WAN acceleration equipment at each branch office. For your private WAN infrastructure (VPN, MPLS or other), most companies are either evaluating, deploying or have already deployed WAN accelerator appliances. These appliances are really changing the IT landscape so they are a necessary technology for most IT environments these days.
So without getting off topic. In our experience, “most” businesses leveraging WAN acceleration technology with branch offices that have less than 60 users can get away with only installing a T1 private WAN connection. Even with VoIP and other big apps being pushed out to the edge. Most host locations or disaster recovery locations will usually require NxT1 or partial to full T3 connections.
Please understand that there are a ton of variables here that relate to:
-types of users
To break it down just follow this checklist:
1 – Baseline your network so you know what is applications are running on it and how much bandwidth each is using.
2 – Identify critical applications and determine the bandwidth need for adequate performance for each.
3 – Identify the “trash” applications such as “Weather Bug”, internet radio, etc. which can be limited to little or no bandwidth.
4 – Filter your outgoing as well as your incoming traffic on all firewalls.
5 – Write ACLs for your routers or L3 MDFs and which log violations to you management console and analyze the results.
6 – Instigate QoS
7 – Build a Lab that represents the hardware and software infrastructure of you LAN/CAN/MAN/WAN/WLAN and test new applications to verify how they will effect you current bandwidth configuration.
8 – Be proactive as you monitor the bandwidth use and it is growing at a more or less constant rate recommend and add additional bandwidth before you create a bottle neck.
Armed the information above …. and with patience and focus ….. you should arrive at a solid answer for the best bandwidth solution for your network. If you would like free help with the process…..just let us know.