Atlantic.Net Blog

Study: Cloud Beats Dedicated Systems on Availability

Availability is one of the biggest concerns of information technology chiefs. The NIH ran a study comparing availability of cloud and dedicated machines. Cloud won.

  • Availability Among Top Three CIO Concerns
  • Availability: Cloud Hosting vs. Dedicated Servers
  • Cloud for Fast Processing of Huge Datasets

Availability Among Top Three CIO Concerns

Unfortunately for CIOs, there are many aspects of their role that can be stressful. For a survey featured in CIO magazine in 2015, 276 CIOs and other top IT leaders discussed the elements that can give them the most trouble; and the top three were security, availability, and making the right hires.

Let’s look specifically at the issue of downtime; in other words, the need to optimize availability.

Availability is second only to security among the top worries of respondents, the poll found. Although business continuity and DR plans can be relatively costly, more than 2 in 5 CIOs (42%) said that disaster recovery is critical for their enterprise and was one of the top priorities for spending in 2015. Those polled said that they think availability is a worthwhile way to invest (via continuity and DR tactics); because if downtime does occur, the public image of the company is tarnished.

Continuity and DR are not things that a business should underfund, explains LifeShare Blood Centers CIO Ric Jones. “So much of budget planning for these services comes down to trust between a CIO and a CEO and others in the C-suite,” he said. “There must be open and honest communication between all the parties involved so when we go to other executives they understand the absolute necessity of these services.”

Beyond the way in which downtime inconveniences an organization, it can also be hugely detrimental to a company’s ability to bring in sales – particularly in fields such as hospitality, according to Starwood Hotels CIO Martha Poulter.

Availability: Cloud Hosting vs. Dedicated Servers

As indicated above, availability is a major concern. How, then, is it best optimized?

Cloud providers have often claimed that their systems offer greater availability than do legacy systems, simply because the network of computers is so vast and redundant. Computer scientists working in the National Institutes of Health lab wanted to know for sure whether cloud could outperform its own on-premises high-performance cluster. They tested the two in direct competition.

The test was for a project headed, in part, by Dr. Maria Mills, a postdoc in the National Heart, Lung, and Blood Institute (NHLBI). She and colleague Dr. Keir Neuman are exploring how small molecules impact a protein. When the researchers started this project, they were specifically curious about whether a molecule would cause any alterations in the protein’s composition or behavior, noted Dr. Mills. “We do an atomistic complex simulation where every part of the protein and every part of the molecule—the physical equation of them—are all explicitly calculated,” she said. “It’s a massive amount of information that’s being calculated very quickly over and over again.”

In early 2016, Alastair Thomson, the CIO of the NHLBI, asked Mills to try an experiment for the next two months that would make use of Biowulf, an optimized Linux cluster with more than 20,000 processors maintained in the agency’s own datacenter. Biowulf was to be compared directly to public cloud. Interestingly, Mills said speed of the two systems was about the same but that cloud use was generally more efficient.

How is cloud more efficient? You can instantly access a cloud system with the parameters your need, but in order to use Biowulf, you have to reserve the processors ahead of time. Sometimes, Mills noted, you had to wait a day or two. “And since I tend to use a lot of processors for what I do—256 to 512 processors—sometimes I wait quite a while,” she said. “And honestly, I sometimes feel like I’m hogging [the processors] a bit, because my jobs can take 60 hours.”

The protein system Mills is studying helps disentangle DNA, which is important: if DNA gets mixed up, it can cause birth defects, boost the risk of cancer, and increase the rate at which the body ages. The researchers affix a magnetized bead to a segment of DNA. They are then able to adjust it with tweezers. The proteins under scrutiny modify the structure of DNA, make cuts in it, and otherwise modify it. The tweezers allow the scientists to gauge exactly what’s happening.

Mills had a significant need for processor power, so it made sense that Thomson wanted to look into using an outside provider. When testing, Mills worked with the cloud host’s personnel to create the best design and to incorporate code that would accelerate the process.

In the end, Mills was a fan of cloud. “[I]t was resources temporarily dedicated to me, so I didn’t have to worry about waiting for them and I didn’t have to worry about other people not having access to resources they needed,” she said. “It was a virtual computer cluster that I built and I’m using, and when I’m done with it I just shut it off.”

Cloud for Fast Processing of Huge Datasets

Are you considering cloud for your next project? At Atlantic.Net, all of our Cloud Servers are powered by enterprise solid state drives (SSDs), providing blazing-fast performance for your servers and applications. Become a cloud hosting member for cloud servers that exceed your expectations, or find out more about our colocation hosting at our Ashburn data center.

Get a $250 Credit and Access to Our Free Tier!

Free Tier includes:
G3.2GB Cloud VPS a Free to Use for One Year
50 GB of Block Storage Free to Use for One Year
50 GB of Snapshots Free to Use for One Year