 |
||||
|
|
|
||
|
Contents 
Introduction
Packet shaping/bandwidth compression
Observer 10.1
AppManager v6.0
Lightspeed Systems Bandwidth Management
Specifications
How we tested
About RMIT |
|
||
|
|
||||
|
|
||||
Packet shaping and bandwidth compression devices allow you to shape and compress data so you can lower your bandwidth costs. They also increase your network performance. Networking vendors like Cisco and Nortel offer routers with queuing capabilities and for someone who requires simple queuing or prioritisation, this may be all they need for basic shaping. However, when there is a need to distinguish specific applications within the traffic, routers will not be able to do much for you.
This is where dedicated packet shapers come into it as they can offer better policy enforcement than traditional network switches and routers.
Packet shapers primarily classify network traffic into categories based on applications, protocols, subnets, and URLs. They can match a user's IP address to specific applications like SAP or Kazaa.
There are many ways you can shape traffic; one way is capping traffic. For example, you can reserve a minimum of 20 percent of the WAN link for e-mail and set the virtual link to exceed the minimum if bandwidth is available but cap it at 50 percent of the link.
You can set all sorts of minimum and maximum bandwidth rates for specific applications. This way you can ensure a smooth and even flow rate that will maximise your throughput.
Then there are compression devices -- these will allow you to compress data, which can accelerate your traffic and minimise your network bandwidth costs, as you will be transferring less data.
These devices can also do some shaping but don't do it as well as dedicated packet shapers. Peribit for example make bandwidth compression devices that use pattern matching technology to compress data. They call it Molecular Sequence Reduction (MSR). The following table shows results on a variety of network traffic that uses MSR. This table was sourced from one of the white papers we found on MSR.
| Application | MSR Data Reduction |
| 75% | |
| Exchange | 79% |
| MS SQL | 80% |
| SAP | 73% |
| Telnet | 67% |
| Oracle | 78% |
| File sharing | 81% |
| Lotus Notes | 78% |
| HTTP Servers | 96% |
| Search Servers | 84% |
| Image Servers | 70% |
During the first transmission a bandwidth compression device will significantly reduce data and after that it starts matching traffic patterns to even further reduce the size of the data. The figures used here are quite impressive to say the least, however the results you get from the device will mainly depend on the type of data you are compressing. For example, if you were to transfer executable files over the network you probably won't be reducing the file sizes by much since the files are already packed very tightly, the same can be said about Zip or MP3 files. If you are interested in reading some white papers on MSR you will find plenty on the Peribit Web site.
