F5's Lori MacVittie recently wrote a piece for ZDNet on cloud congestion and I thought I'd summarise it for our UK TechView readers:
As cloud-based application usage continues to increase among consumers and corporate employees and more businesses plan cloud services without considering the ability of internal networks to cope with the extra traffic, cloud congestion is becoming more and more of a likelihood.
It is the failure to increase bandwidth at the corporate level to withstand the onslaught of traffic before embarking on a large-scale cloud initiative that will cause problems.
It's true that web requests and responses are generally not big data and are not as time-sensitive as voice and video. But there will be many more of them, and lots of little web requests and responses traversing the corporate internet lifeline will have as big an impact on the performance and capacity of that network as the combination of voice, video and email.
As WAN links become choked with traffic from a variety of business and — admit it — non-business uses, performance suffers. Congestion-avoidance techniques and WAN optimisation can provide only so much relief before it becomes obvious the only answer is to increase the number of lines or upgrade to a fatter, faster pipe.
It is critical to the overall success of any cloud migration project that the capacity of the network be evaluated to ensure it does not become a bottleneck. Understanding the impact on the business of productivity losses from performance-impaired networks is paramount in building the business case needed to justify the upgrades to the network before launching the migration.
Don't be fooled into ignoring the network when the reality is that migrating to cloud means you're making the network a focal point, the single point on which much of your business will rely.