Story image

Overcoming WAN pains

01 Mar 11

A WAN is something that will be considered eventually for any company that has more than one office. Often a fully-fledged WAN will grow over time, and may start out as a simple VPN tunnel over two public ADSL connections, however once a company is up-and-running with an ad-hoc WAN, growing pains may start to occur.
If data overage on your ADSL plans is a common theme, or you’ve got three different IT companies looking after five different pieces of equipment, in four different locations, then you may be experiencing the kind of WAN pains that make you wonder how a WAN solution can be affordable and useful at the same time.
A WAN might seem like a risky or unnecessary investment, and staff may simply go back to emailing files between remote offices to get by. However, deploying a properly and sensibly thought-out WAN solution can save money and increase opportunities in a number of areas. Deploying solutions that have a short ROI, while delivering increased functionality, productivity and performance gains, is a strategy that makes sense.
I believe that for most organisations, the primary reason for implementing a WAN, or WAN optimisation, will be to see an improvement to the bottom line. This assumption follows from the fact that shareholder returns are often the primary focus of an organisation set up along standard capitalist lines.
The improvements to the bottom line may come in the form of reduced costs, improved staff effiency, or a security crisis that has been avoided, but regardless of how they come about, it is worth mentioning that you should be able to see the financial benefits of implementing a WAN or WAN optimisation before you start. Don’t simply get pushed into implementing a WAN by your IT advisors, if the benefits aren’t tangible to your organisation.
With a properly configured WAN, you get dependable, guaranteed circuits between locations, and in many cases are able to decommission existing circuits.
A WAN often allows you to consolidate all facets of your IT infrastructure, thereby avoiding some of the whole cycle of router upgrades, server refreshes, backups and redundancy for stand alone offices. A big plus on this style of deployment is that it instantly future-proofs you for many projects, and opens your organisation up to be able to offer a wider scope of products, or better service and support, as all your sites are interconnected. Essential ERP, CRM and Infrastructure services such as voice and email can be quickly deployed and centrally managed.
A multi-site deployment via a WAN infrastructure also gives any company the ability to provide a true disaster recovery scenario. As many Christchurch business can attest, having the ability for business to continue, even if one office is unable to function, can prove the difference between solvency and insolvency.
Rather than having to outsource data backup solution specialists, an in-house solution can be developed that means the business has a true failover if something should happen to one of the branches.
The ROI on WAN deployments is often fast, because initiatives can be deployed quicker and provide cost savings in the form of increased mobility of works, with less distributed infrastructure required, whether that is voice, data or email. Telephony costs that might have included a toll call to another branch are suddenly an internal call.
You can focus more on your core business and less on the peripherals of maintaining a distributed infrastructure. You can convert sunk costs and unpredictable IT expenditure into something that can be forecasted and controlled better, allowing you to scale dynamically as required.
Productivity increases, workers are happier that the performance of their applications has improved, allowing them to be more satisfied with their job. Slow performance and frustrating links issues are a major productivity drag on any business, especially when links go down. All these existing bottlenecks that happen in Site-to-Site VPNs, over the internet links and via remote workers, disappear with the implementation of a WAN deployment.
Cloud based services are another avenue, and reason for implementing a WAN infrastructure.  Whether you go with an outsourced provider, or do the virtualisation in house, WANs provide a LAN-like connectivity to services, making the virtualisation more successful and end users maintain the application performance they’re expecting.
A critical part of most businesses these days is IT support, but with a multi-site deployment, this can be problematic if you’re not dealing with the same company for support in all the locations. With a successful WAN deployment, support and troubleshooting that previously might have had to have been undertaken by a third party can be brought in-house and dealt with by staff.
Workloads can be balanced better, through centralisation of services. Transparency improves as more can be managed by a smaller team and there’s no reliance on third-party service providers and IT vendors to manage remote sites. This eliminates significant cost in travel such as flights, hotels, and transport. Staff can actually work on the project without being torn away to travel to remote locations.
If you choose to outsource critical services, such as server hosting or voice, a WAN means LAN-like performance to that vendor without internet connectivity being involved or compromised. Once a WAN has been implemented, various optimisation strategies can be employed to ensure that the connections run in an optimal way. These strategies may include:
Simple Rate Limits: These can prevent individual users, applications or sites from congesting a particular link, and ensure quality of service for more of the time.
Traffic shaping: More intelligent soultions than simple rate limits can be introduced, and there are solutions from Cisco, Packeteer etc, that will allow you to prioritise voice traffic, high quality data or other applications to make the best use of a link.
Latency optimisation: A WAN optimiser can be installed that will manage latency across links, and some can even answer queries locally, and then forward those queries on to a remote site in order to reduce the appearance of the delay of the WAN to an end client.
Protocol spoofers: These can bundle multiple requests from a ‘chatty’ application into a single request, thereby minimising re-transmits and magnified latency across a link.
Forward Error correction: This strategy can minimise packet loss across a link by introducing additional packets to reduce re-transmits in error-prone and congested links. Although it is a sophisticated strategy to implement, a small percentage of packet loss can have serious impact on a high speed link.
Just 0.1% packet loss on a 1Gbps link can potentially reduce throughput to as little as 7Mbps depending on the application. People can really under estimate the impact of packet loss, and therefore end up paying much more for the physical links than may be necessary.
Caching and proxying: Where the same data is accessed over and over again, a cache in each remote office can keep a local copy and drastically reduce requests on a link.
Duplication of data: Data can be duplicated across multiple sites for local access, and then syncronised at off-peak times, such as the early hours of the morning. This may work for data when a single ‘master version’ is not required.
A WAN can be built out of a combination of ADSL, Wireless, Fibre and leased line infrastructure, and the success or failure of the network is often simply down to planning and implementation.
Regardless of which ISP or telecommunications company you are with, there are many organisations that can assist in trouble-shooting a WAN implementation, in order to get it running the way it should, including Voyager. If your WAN isn’t working, seek help!