IT Brief New Zealand - Technology news for CIOs & IT decision-makers
Story image
WAN optimisation: what a difference a layer makes
Tue, 10th Apr 2012
FYI, this story is more than a year old

The application layer of the Open System Interconnection (OSI) stack has long been a focal point for Wide Area Network (WAN) optimisation. Efforts to optimise application delivery have been centred here because the pain was rooted in optimising only a handful of applications at branch offices.

However, there’s a revolution happening today in WAN optimisation (WANop) where projects like data center consolidation, cloud computing and virtualisation are putting the application-based WANop approach on its heels.

Today there are far too many applications, platforms and versions of applications to account for. This is compounded by the popularity of lower cost multiprotocol label switching (MPLS) WANs that are ripe with bandwidth limitations, latency and packet loss.

The seas of WAN optimisation are beginning to part, and IT organisations must not only deploy WANop to recoup valuable bandwidth for specific applications, but deploy a solution that optimises the entire application environment today and in the future.

So, why would anyone use a WAN optimiser that addresses only some applications but not others?

Historically, purchasers of WAN optimisation products most likely believed or misunderstood that the product they purchased was optimising the entire stream, when in reality, it was only optimising specific pain-point applications and bypassing all others. But what choice did they have? Most WAN optimisers only provided capabilities to optimise specific applications.

To understand this further, one needs to look at underlying protocol support. While the lion’s share of network ‘pain’ traditionally has been based with TCP applications, the number of business applications that are non-TCP based is rapidly growing. Voice and video come to mind, as well as desktop virtualisation (VDI) and even home-grown apps. If a particular WAN optimiser doesn’t support the non-TCP apps, it is by design forced to bypass that traffic. This alone should set off red-flags.

But for application centric WAN optimisers to address this, the answer isn’t so simple. It means these vendors must apply ‘plug-ins’ for every application the customer wants to optimise. This is not only complex because of the ever growing application set in today’s enterprise, but also presents challenges because of the dynamic nature of the applications. In short, if the software or application vendor makes a change to the app, the app-centric WAN optimiser must update its plug-in to support it.

Enter the network centric WAN optimiser, which does not care if applications change or get updated. Network centric WAN optimisation is handled at the network layer (layer 3). This means no plug-ins are required for optimisation and ensures any new applications added to the environment are optimised with the existing WANop investment.

An average enterprise can have as many as 80 applications—everything from ERP to unified communications, to backup and data replication. Some of these might be purchased from third-party vendors, some might be hosted in the cloud, while others may be proprietary. If the enterprise relies on the app-centric WANop approach, the time, cost and effort to ensure optimisation is working properly for each application could be astronomical.

Then there is the network itself. MPLS and shared networks are growing in popularity because of the low-cost and flexibility. However, these popular WANs are plagued with network quality issues that adversely impact applications.

This is another area where the network centric WAN optimiser has the advantage. Advanced WANop technology such as Forward Error Correction (FEC) can operate only at the network layer to improve application performance. This technology can reconstitute lost packets in real time to avoid delays that would otherwise come with round trip retransmissions required by application-based WAN optimisers.

Cost-effectively optimising all applications

Of all the reasons for performing WAN optimisation at the network layer, application independence is probably the most obvious. Such a network centric WAN optimisation approach enables organisations to future-proof their networks and allows for customers to get a better handle on trends such as cloud computing, virtualisation and data center consolidation.

Network-based WAN optimisers provide the advantage of optimising all current and future enterprise applications. Whether enterprises update to new versions of their applications or update existing solutions, by default, network WAN optimisers continue to operate in lock-step with the dynamic application environment

Wayne Neich is regional director, Oceania for Silver Peak Systems. This article originally appeared in the April issue of IT Brief; go here to subscribe.