How to Accelerate Your Cloud Service

By Brian Boguhn
Updated: May 24, 2012

The push these days is for businesses to move to the cloud. Yet for all of the hype the move to the cloud has received, there remains one issue that most companies are failing to get a handle on: latency. There are a variety of causes for latency. The distance between the office and the data center can be one. There can also be congestion on the network along with packet loss. Is there any way to accelerate things to lower the latency and get better performance? This article will discuss several options that can help.

Distance

Many times companies have their offices located in one part of the country and their data centers located in another part of the country. In the United States, a company can be based on the west coast but have its data centers on the east coast. This distance between the two can lead to slow file transfers and inefficient application performance. For some applications, it is critical that the cloud provider have a data center that is in the same region as the company’s office. Choosing a provider with multiple data centers across the country can help in doing away with the latency caused by physical distance issues.

Congestion, Packet Loss, and Windowing

Other key issues that can contribute to latency are congestion on the network, packet loss, and windowing. Congestion affects latency by holding some data back in a queue when the total amount of data to be sent down the connection is too large. Windowing is caused by the way signals check that they have been correctly received; the receiver must send a message back to the sender that it has the first packet sent before the sender will send the next packet. Obviously, this means more delay and more congestion for every packet. If packets are lost during the transfer, the problem only intensifies as replacement packets need to be sent. The speed of travel from one point to another slows significantly with all of the queuing, confirming receipt, and resending.

The traditional methods of WAN acceleration or network Quality of Service (Qos) are not well suited to businesses trying to prioritize cloud-based services from public providers. As a result, companies are seeing that it is very important to consider cloud service providers who can offer application performance guarantees as part of their service and providers who can be measured and backed up by stringent Service Level Agreements (SLAs) when entering into contracts. This is the only way to ensure that users do not see 10 to 20 second service delays when using business critical applications on a day to day basis.

Monitoring

The final step in accelerating your cloud service is to incorporate monitoring. Once the issues of distance and latency have been addressed, the service provider monitoring of end user activity in the cloud can help to minimize capacity shortages and bottlenecks. There are other more sophisticated tools that can provide continued analysis of service performance available to the IT team tasked with managing the performance and capacity of cloud provided services. These tools help the IT team to make sure that the delivery to their business of the promised benefits of the cloud is actually occurring.

Conclusion

The cloud offers benefits that businesses can now take advantage of that previously were not available. With these benefits, however, come new challenges that cannot be addressed in ways previously used. By addressing physical location and congestion on the network, as well as introducing proactive monitoring, businesses don’t have to suffer with cloud service that doesn’t deliver on its promise.