All news

Getting Control Of The Cloud - How Better Cloud Management Can Prevent Risk & Deliver Crucial Infrastructure Insight

June 13, 2009

Despite its name, cloud computing appears to offer clear, not cloudy, advantages to data centers because of its ability to deliver impressive computing flexibility and massive capacity without the need to buy equipment. But as enterprises jump headfirst into this new infrastructure, some are discovering that life on the cloud isn’t necessarily transparent.

Despite its name, cloud computing appears to offer clear, not cloudy, advantages to data centers because of its ability to deliver impressive computing flexibility and massive capacity without the need to buy equipment. But as enterprises jump headfirst into this new infrastructure, some are discovering that life on the cloud isn’t necessarily transparent.

Research from IDC indicates that worldwide IT spending on cloud services will reach $42 billion by 2012, while Gartner is even more bullish in its forecast, predicting that worldwide revenue from cloud services will hit $56.3 billion this year. What might have been disregarded as fancy hype just a few short years ago is now seen as a legitimate business model, with managers recognizing the cloud as a reliable, long-term solution to getting spiraling costs under control.

Yet with the cloud still in the infancy stage, IT organizations are often left struggling to determine precisely how the cloud interacts with the data center. Traditional methods of solving problems with applications, services, and other onsite elements won’t necessarily apply when cloud services are thrown in the mix, so the need for finding and deploying more effective means of managing the cloud is quickly becoming apparent.

Storms Lurk In The Cloud

According to Pete Beckman, director of the Leadership Computing Facility at the Argonne National Laboratory, cloud computing offers a low barrier to entry, which is particularly crucial in the current economic climate. “Being billed for success—customers using the infrastructure—instead of paying for what you think you might need is a very appealing model to CFOs. Nobody likes gambling on overprovisioning,” Beckman says.

But there’s a flip side to riding this seemingly effortless new wave of hosted services, and it might not be readily apparent to data centers eager to take advantage of its benefits. In terms of flexibility, cost, time, and complexity, the network infrastructure faces new requirements to support this model, explains Jarrod Siket, senior vice president of marketing for Netronome Systems (www.netronome.com).

“The network that lives between the user and the host content, application, or service must be capable of delivering an experience to the user that makes their separation from those resources completely transparent. This can fail, and enterprises can suffer not only from outages or slowed application performance, which affect productivity, but from a host of security issues related to threats introduced via the cloud,” Siket says.

Know Your Cloud

Most enterprises consider knowledge as power. But for IT organizations, knowledge has become risk mitigation, according to Patti Dock, COO of DataMotion (www.datamotion.com). Cloud services might present themselves as infinitely reliable, but Dock stresses the need to perform proactive due diligence on vendors and their operations to mitigate risks. For example, how is the information processed? How is it transported? Where does the data reside? What visibility into what’s happening is available?

“Reporting, monitoring, and tracking in real time or near-real time should be part of the deal,” Dock says. “Governance of information is critical to properly managing a small cloud deployment to a huge interdependent application architecture. Many enterprise cloud applications integrate with other on-premise or SaaS applications. Dependencies can get a little murky when various applications from various providers start hooking together to create a business process, and customization becomes a big, ongoing piece of the puzzle. Be aware of best practices and choose some cloud key metrics to focus on in order to gauge post-deployment impact.”

The science of gaining insight into cloud dependencies, latency, and other factors is still new and growing, but getting started isn’t difficult or necessarily expensive. And cloud vendors often offer management and monitoring tools with their offerings.

“The amount of control enterprises may have over [cloud] issues will differ depending on how they’re using the cloud and how interconnected their network is with the cloud. As long as enterprises have dealt with these issues internally—especially those dealing with securing their network—and done their homework to make sure that their vendor meets requirements, they [will] better manage these dependencies and deal with problems that may arise,” Netronome’s Siket says.

Back To Process

 

Argonne’s Beckman says that cloud services require a potential alteration to the traditional IT mindset. Perhaps not surprisingly, this change falls in line with the similarly growing trend of melding business with IT.

“I think the first step for enterprises is to take an inventory of what they want to do, not what software they have,” he says. “Successful paradigm shifts must be driven by business goals, not technology. Moving to SaaS or IaaS [infrastructure as a service] might also mean changing processes. Once the processes are well understood, the dependencies can be addressed.” 

Christian Perry