Your WAN capacity is running close to its limit, users are complaining about slow loading speeds of business critical applications and backups are falling outside of their windows... this scenario is played out in companies across the world, so how do you solve the problem?
Reducing the raw data travelling across your WAN by either blocking access to certain bandwidth heavy websites or deploying a WAN Optimiser can take the pressure off for a while, but you will still be at risk of suffering slowdowns and high latency due to the steady increase of your base traffic levels or from bandwidth hungry applications flooding all the available capacity. It’s obvious another technology is needed to guide and shape the bandwidth in addition to optimisation and acceleration, making sure specific applications and protocols either have access to a minimum level of bandwidth or are limited to a maximum level and can’t take all of the bandwidth for themselves.
I am, of course, talking about Quality of Service (QoS).
On its own, QoS helps to partition up your WAN link into many “virtual lines” and will throttle or burst traffic depending on your needs and the configuration set on the device. The drawback here is when you don’t have any WAN Optimisation you can quickly outgrow your connection because you’re trying to squeeze traffic down an even smaller section of your pipe.
Taking advantage of QoS usually means replacing network infrastructure with expensive switches, choosing your firewall because it supports that feature or even buying a device which is dedicated to the task. WAN Optimisers from Sangfor includes QoS as one of its core features for bandwidth management and together with traffic optimisation/acceleration make a complimentary pairing to get the best performance out of your current WAN infrastructure at a fraction of the first year upgrade cost.
This works because our units can identify traffic based on its TCP port and by using Layer 7 Application Identification, apply optimisation to the traffic to reduce the flow of data and then pass this reduced traffic onto the QoS module in order to shape it in accordance to your specific policies. Since we can identify traffic at the packet level, you don’t need to set up any additional rules should it be running across a non-standard ports – such as VMWare server using HTTP across port 8333 for its administration interface – reducing the maintenance required on your policies whenever you implement a new system/service.
Through this unified approach, not only can you dramatically improve the performance of your WAN communications but you can keep a constant eye on your changing usage patterns and respond based on solid data provided by our reporting module. This way you won’t ever let a single application bring your WAN to a crawl or waste a single byte of your precious bandwidth.
When thinking about WAN Optimisation, it is easy to concentrate on raw bandwidth figures or noticing the time difference between backup jobs from pre-optimised and post-optimised days - Epochs? Eras? - and miss another vital aspect of improving the efficiency and performance of your networks – understanding who is doing what and with what across your WAN by making good use of the reporting features present in the WAN Acceleration and Optimisation units from Sangfor.
By knowing which machines and what services are generating the most traffic and at what time of day, you can diagnose the root cause behind incidents and adjust the settings of your Sangfor unit to better suit your needs and provide a solution, especially when adhering to ITIL best practices.
For example, you have installed a new WAN Acceleration and Optimisation unit into your head office and a branch office but Remote Desktop users still report periods of noticeable slowdown. As part of your investigation, you can run a report covering the time scales of the slowdown to see which IP on your network was generating the most traffic and the services that traffic was composed of.
Through that report, you may discover a server running a DR backup/replication or a particular user making heavy use of media streaming websites resulting in less bandwidth for Remote Desktop users. Armed with this knowledge you can immediately address the issue by either making changes to the schedule on the problematic server, limit the user from accessing those websites or by modifying the Bandwidth Management settings to ensure essential traffic such as RDP is guaranteed a minimum slice of the bandwidth pie, reducing the impact other services have on a business’s core applications.
Where you could have spent days collecting and reading TCPDump logs or spending money on extra network monitoring equipment and contractors, you can generate a report in minutes, pinpoint the cause and develop a solution within a very short period of time. By setting these reports up to generate automatically and be delivered straight to your inbox on a daily, weekly and even monthly basis you can often notice a problem before it becomes a disruptive one.
In addition to being a useful tool in diagnosing issues in the present and formulating a solution, this body of data can assist long term projects in making the case for or against any major changes to the provision of IT services across the company. Traffic Flow Speed and Total Flow reports will show just how much data is passing across your WAN and how close you are to the maximum throughput your WAN connection is capable of.
This allows you to answer important questions such as “Can our infrastructure handle the migration of our CRM application to a new cloud based solution?” or “Is the bottleneck our WAN link or the server performance?” which often leads to “How do we fix this?” and, more importantly “How much will it cost?”.
Taking advantage of the built-in reporting functionality in Sangfor’s WAN Accelerators gives you an in-depth knowledge of the data flowing in and out of your network, which is crucial in providing an efficient WAN, ensuring your business gets value for money on their infrastructure and informs future decisions based on solid usage data.