WAN optimisation has been typically deployed in enterprises to recoup precious bandwidth and speed up sluggish applications. But, now with the growing list of applications, platforms and versions of applications, the problem has become boarder; it is no longer about optimising only a handful of specific application at branch or remote offices. The advent of cloud and virtualisation has further exacerbated the problem, making it an imperative to deploy a solution that cost-effectively optimises the entire application environment now and in the future, and do it without increasing network infrastructure complexities.
The primary function of the technology – making response times faster over WAN links – now available through software that runs on virtual machines, is becoming practical for use in public and private clouds where virtual environments are the norm.
The number of applications that can’t run in a virtualised environment is rapidly dropping and WAN optimisation vendors are delivering virtualised support for their applications, further reducing the need for stand-alone appliances. Running WAN optimisation on virtualised servers enables IT to use the same servers that it uses for other essential branch services (DNS, DHCP, print services, etc.). This eliminates the need for a separate box in each location, if there is already adequate hardware in a typical branch, and thereby reduces administrative costs and reduces the weight of the deployment. Virtual appliances are easier and faster to deploy, and to move, and they offer greater options for resilience and disaster recovery.
Virtualised versions of the old hardware appliances make it possible to deploy optimisation within public cloud provider networks, meaning cloud-based applications respond better. It also means data can be sent in less time to cloud storage facilities where it occupies less disk space (and so costs less) and is secure because it is encrypted.
So how does WAN optimisation fit in the cloud space?
“As applications and storage move out to the cloud, performance of the corporate communications link between the data centre and the cloud will become more important. Optimising transfers to the cloud with symmetric WAN optimisation technology improves WAN communications performance and helps contain costly bandwidth upgrades,” says Diego Arrabal, sales director, F5 Middle East.
Florian Malecki, senior product marketing manager at SonicWALL, echoes a similar opinion: “WAN optimisation is definitively playing an important role with cloud computing, as repetitive data are exchanged and internet bandwidth becoming saturated can be expensive. Organisations can leverage and optimise their existing connectivity with solutions.”
There are several aspects to the decision-making process surrounding optimising application delivery in the cloud. One involves whether to apply a symmetric or asymmetric solution. Asymmetric solutions are those that require hardware on one end of a connection, while symmetric solutions require gear on both sides of the connection. For many Web-based apps, asymmetric offerings are the best approach to accelerate applications from within the data centre.
Both asymmetric and symmetric approaches are available as hardware, software and services, and vendors says asymmetric may not be the best option for cloud environments. “This would only work if the data and apps held in the cloud were static. If you want to make the cloud a true extension of the data centre with business policy applied at both the data centre and the cloud, then you need the means to apply this policy and that means having application delivery controllers, either physical or virtual, at both ends,” says Arrabal.
Dr Steven Turner, Optimisation Consultant, Intergence Systems, agrees,“ Typically, acceleration should be symmetric to enable a user to transfer data to the cloud as well as receive data from the cloud. Looking at cloud based applications such as Sharepoint there is a definite need for two-way acceleration to support both the transmission and receiving of documents and other media.”
Hardware vs. services
The two best things about WAN optimisation are that it practically guarantees better response times for applications, while at the same time either reducing the need for WAN bandwidth or at least staving off for a while the need to boost it.
Initially the technology was deployed via appliances placed at both ends of WAN connections that perform a variety of optimisations that reduce the number of bits that have to cross the wire in order to complete transactions. But over the years, providers have stepped up to offer WAN optimisation as a service.
Some providers offer optimisation services based on installing appliances made by somebody else at customer sites and managing them. Other providers offer optimised networks over which customers run their traffic with no need for devices at customer sites.
Arrabal from F5 says there is also potential to offer WAN optimisation as a service over the cloud. “It could potentially make up part of an SLA or service offering to cloud clients – the benefits of bolting on such a service could include better network access to applications, reduced bandwidth costs etc. But ideally you’d like to see this kind of service being offered as a way of delivering apps better and optimising the WAN as a matter of course,” he says.
WAN optimisation appliances constitute a growing market now, but the writing is on the wall: A few years from now “WAN Optimisation Controller” may barely exist as a separate category of physical network appliance. As WAN refresh cycles play out over the next seven years or so, the research firm Nemertes expects a majority of companies to migrate to one of the other optimisation models now available: carrier/cloud services (optimisation as a service), point solutions and virtualised optimisation appliances (optimisation as a workload), or integrated router-switch/optimiser or security/optimiser appliances (optimisation as a feature).