enVirtua.com

Your organisation is distributed across offices, connected via your Wide Area Network or WAN. You might have the fastest servers in the world, but if the WAN link is not proiding the service you need, things will run slow.

There are many ways of optimising WAN traffic to improve performance. You can throw more bandwidth at the problem or you can move servers out to your remote offices. These are the two most common solutions and also the two that deliver the most complexity and the most cost often.

There are better ways to optimise WAN traffic!

Traffic Shaping

Rather than allowing TCP/IP traffic from all your applications, emails, and web traffic to compete for the same limited resource; you can manage which types of traffic have priority and how much bandwidth they are allowed to have.

For example, SQL Database queries are typically time sensitive. You do not want your point of sale (POS) or stock control application having long delays as it waits for an answer from a central server. Especially if that delay is being caused by someone watching Big Brother on the web!
With traffic shaping we are able to specify that SQL traffic is of more importance than WEB traffic. If a user makes a SQL request, it will be given priority over the web traffic. It will slow the web traffic as it gets queued beind the SQL request, but what matters more to your business?

Another common example is Citrix traffic. Citrix requires a good amount of bandwidth and is very time sensitive. If your WAN gets congested with say emails with attachments, then users desktop environements will suffer slowdowns.
What is very common is to allocate each Citrix session a set amount of bandwidth and to gurantee that amount of bandwidth. We would also set the Citrix traffic to a high priority. The result is that in effect each user has a dedicated “pipe” which has enough bandwidth as required. That “pipe” is also set to go across the WAN before say emails which there is little time sesitivity to.

Of course there is a exception to every rule. Recently it has become common for businesses to prioritize email above other traffic. Specifically emails outbound to Blackberry devices. This is an example of how each WAN needs to be carefully assessed and managed in a unique configuration. Sure there are commonalities, but each WAN has different requirements and needs to be monitored and modified on a regular basis to make best use of the scarce resource than is network bandwidth. Which leads us to:

Traffic Monitoring

Closely linked to traffic shaping and integral to it is the process of network monitoring and assessing your network traffic. The trick with traffic monitoring is being consistent and applying changes to your traffic shaping in an interative manner. Monitor -> Make a small change -> Monitor -> Repeat.

What often happens is that a WAN optimisation solution is put in place and then is left running. It is not monitored and tweaked, and eventually network problems occur and the investment in traffic shaping or bandwidth or some other solution is put into question.

Automated monitoring and reporting is always the best way to keep an eye on your WAN traffic. If that is not feasible then regular planned manual reporting is possible.

In either case, the important factor is NOT the monitoring or reporting, rather the actions that the reports cause. Part of the reporting/monitoring solution you put in place needs to be making alterations to your WAN optimisation configurations.

Wide Area File Services (WAFS)

Managing the traffic that goes across your WAN may well solve your network performance problems. But often it just makes the problem smaller. You go from long delays to medium delays. WAFS, or Wide area File Services, can solve that problem by providing remote data access at local network speeds.

Typically a WAFS solution will optimise your WAN traffic in a number of ways to obtain the best performance possible. below are some examples:

• Caching.
  Normally each time someone opens a document, lets say a standard sales letter, it is pulled across the WAN. With caching, the file is pulled once and then if it is needed again the file is served to the user from a cache stored in their location.
• Protocol Optimisation.
  The TCP/IP protocol used most commonly today is what we describe as “chatty”. Which means that when a file is transfered across the WAN it may take thousands of small packets to and fro. SYN and ACK packets fly backwards and forwards, all delayed by the congestion on your WAN and the round trip time between sites.
  This can cause large delays, which can be reduced by optimising the file transfers at a protocol level. So changing packet sizes or even using alternative methods to TCP/IP.
• Data compression
  In very simple terms you could imagine this like having a box that compresses every piece of data for you before you send it. A bit like ZIPing a file before you email it to a colleague. Except that this data compression can happen to all your data regardless of type. And it may be able to do compression across files.
  So for example, you company logo is probably embedded in multiple files. Each time one of these files is opened the logo is sent across the WAN. If that specific file is sent again, then caching would prevent the logo being sent. But if another file is sent, the logo goes with it. With data compression that is prevented, the “box” sees that the logo is identical and sends a token tha identifies the logo. The logo is never sent across the WAN the second time, it is just generated at the far end… as if it was pulled from a ZIP file.
• Application Optimisation
  Some applications can be optimised for WAN traffic. For example: CIFS (Windows file sharing), MAPI (Microsoft Exchange), NFS (UNIX/Linux File Sharing), HTTP (Web applications), HTTPS (Secure Web applications), MS-SQL (Database applications), Microsoft Office and Lotus Notes.
  Traffic generated by these (and other) applications, can be modified to improve performance; as each applic ation has quirks that can cause bad performance over the WAN. Application specific optimisation can reduce the application specific issues.

What enVirtua offers:

Your network issues can become our problem, not yours. We would (ideally) begin by implementing some traffic monitoring and regular reporting. This would give us the required knowledge of your traffic to make some informed traffic shaping and optimisation decisions.

Having made some decisions about how to improve your WAN performance, we would proceed in small increments. Making a change and then closely monitoring the trafffic and user experience. Starting with your most important business applications we would find the best performance for these applications first. Of course you need all your applications to work, so we will work with you to ensure that a balance is found between what is possible, what is required and what is not acceptable.

Throughout this process you will receive regular contact from enVirtua. You will know exactly what we are doing, what we are planning and what results we deliver. You can receive updates as often or infrequently as you require. reports can be in person, in writing, by phone or email; it’s entirely up to you.

How to get started?

The best way to get started is to telephone or email us at your convenience.
We can have a relaxed conversation about your WAN and any problems you are encountering. You can talk to us about your ideas for improving the situation and how enVirtua might be able to help.

From this point, we can arrange a face to face meeting, or give you a proposal to deliver to you our WAN services (described above) that match what you need. We will give you an estimate of costs of course.

From this point on we will act as if we were one of your in-house team and work as hard as possible to deliver the best performance from your WAN to your users.