How to leverage WAN optimisation for high-performance cloud services

Cost savings, convenience, speed-to-market and freedom from technology lock-ins are all benefits of migrating business applications to the cloud, and many that have already adopted cloud strategies have indeed realised significant cost savings and productivity gains.

However, some have also experienced performance and availability issues that have led to resistance to migrating business-critical functions. As a result, the bulk of successful cloud projects to date have been departmental and not ‘mission critical’ core systems which have significant requirements for resource-intensive functions such as database replication, file synchronisation, backup and disaster recovery between data centres.

Part of the problem is that the public cloud ‘Infrastructure as a Service’ (IaaS) market is evolving more slowly than other areas such as ‘Software as a Service’ (SaaS). Providers are only beginning to recognise the level to which performance levels need to be increased and the capabilities their offerings need to make them a viable platform for mission critical purposes.

One area of IaaS that providers have quickly evolved is services that support wide area networks (WAN). The WAN is fundamental to any globally connected service offering, dictating data transfer performance between locations. WAN performance optimisation therefore needs to be embedded as part of the cloud infrastructure. By optimising bandwidth and latency times, WAN managers can tune performance and efficiency so that data transfer times are improved and service levels to users are met consistently.

In some situations, WAN optimisation can be achieved by user organisations installing their own virtual client. However, it is far more efficient for the service provider to carry out WAN optimisation between cloud data centres, transparently accelerating traffic for enterprise users. By integrating WAN optimisation into the cloud platform on a global scale, the speed of transfer between locations can be optimised for the WAN as a whole so that resources and bandwidth are deployed where need is greatest and thus overall efficiency is best delivered. Dramatic additional improvements in performance can be achieved by controlling compression and caching within network protocols (e.g., TCP, UDP, CIFS, NFS, MAPI, HTTP, HTTPS, MS-SQL) and databases accessed across the network.

Once organisations are confident that they can optimise the performance of the WAN that interconnects cloud-based applications and data, they can start to migrate mission critical data and applications to the cloud.

Service Level Agreements (SLAs) for infrastructures are generally very demanding for mission critical applications – particularly where they govern up-time, response time and latency performance. And as more resource intensive, mission critical applications make use of IaaS and other cloud offerings, so WAN optimisation becomes a key part of meeting SLAs for performance and availability. As a result, when evaluating cloud services, most prospective purchasers for mission-critical projects will expect core WAN optimisation to be a part of the cloud service and not provided as an add-on with an additional charge.

The use of cloud technologies and services will continue to evolve and take-up will continue to expand. However, for mission critical enterprise-wide applications, WAN performance is a critical issue, and cloud offerings that integrate WAN optimisation capabilities directly into the service will quickly position themselves ahead of less sophisticated IaaS offerings. While those services may sometimes appear to carry a higher initial cost, the performance gains and bandwidth savings that they offer will quickly reduce the overall cost of IT infrastructure.

Keao Caindec is the Global CMO of cloud solutions at Dimension Data