Pagoda Blog


The Expense That Cloud Customers Often Forget

April 22, 2013

By Ken Presti


A move to the cloud is typically a lengthy and detailed discussion that covers a variety of key areas, including cost analyses, infrastructure decisions, security risks, and a host of other elements. But channel partners tend to agree that one very important aspect tends to slip through the cracks.

"People don't factor the cost of cloud bandwidth into the equation as often as they should," said Daniel Cheng, president of AMA, a Mississauga, Ontario-based channel company. "Salespeople should be careful to create the proper expectations about that. Accessing servers that are located off the premises consumes a lot more bandwidth than is normally the case with on-premises solutions. The customer's dependency on the Internet pipe becomes much more pronounced and is increasingly critical to the business. These costs are frequently left off the table at the time that the services are being sold."

Cheng's point of view resonates with Greg Emplit, general manager of Nextpoint Information Systems, a Seattle-based channel partner focused on IT consulting and cloud computing. "People need to pay more attention to bandwidth costs, but they also need to put together a strategy for redundant bandwidth that they can engage if their primary source has an outage. Those are two very important things."

[Related: The Top 10 Enterprise Cloud Services Companies]

Without proper attention to bandwidth when the cloud migration strategy is being developed, Emplit says that customers can soon find themselves faced with difficult decisions on whether to increase their fixed costs or deal with latency issues that will result from inadequate performance.

"Customers can get into the cloud pretty inexpensively, but if the bandwidth costs go through the roof as a result of using the cloud, that can eat into their savings pretty quickly," he said. "There are also challenges in locations where bandwidth is limited and expensive. In some extreme cases, there might not be enough bandwidth to support cloud solutions. When this happens, you can try to find a way to get bandwidth to those locations, but sometimes you cannot do it all."

It's widely known that some extremely sensitive applications are often best left on the customer premises, while the remainder of the IT functionality can be shipped off to the cloud. Emplit added that the cost and availability of bandwidth are among the key factors that should be factored into those decisions.

"We just met with a law firm that has six or seven different applications," he said. "We have to consider which of those applications are well-suited to a virtual environment and which of those applications suffer additional latency when they're hosted somewhere else. And sometimes we find ourselves in situations where different applications are hosted in different data centers, and still other applications are kept on the customer prem. In hybrid solutions like that, we also have to figure out who is responsible for which services."

 

Although industry standards for bandwidth are nowhere near carved in stone, Nextpoint's Emplit advises channel partners to pay close attention to how much bandwidth is required for different applications.

"We don't have a formal benchmark that we use, but for of VDI solution, for example, you probably need about 64K per seat," he added.

Access to bandwidth can also be dramatically impacted by the consumption of other customers who happen to be located in the same multitenant environment. This is an aspect that needs to be factored into the equation, but it is also very difficult for partners to control. These realities can, in effect, decide whether a specific application should go to the cloud under the specific customer circumstances.

"In a multitenant environment you can have noisy neighbors who are sucking up bandwidth," said Jeremy Przygode, CEO of Stratalux, a Los Angeles-based cloud VAR. "One thing you can do to help maximize efficiency is to avoid firing up servers and leaving them on indefinitely. You have to shut them down when you don't need them and add capacity when you do. Automation tools can be of great value in this regard, especially if you're working with large customers who have lots of servers. This is often the key to getting the most cost savings that the cloud can provide."

http://www.crn.com/news/cloud/240153316/the-expense-that-cloud-customers-often-forget.htm?cid=nl_crn&elqTrack=true




Return to Pagoda Blog Main Page