Virtualization, cloud streamline datacenter operations, costs

Beyond reducing need for more physical servers, storage and capital outlay, cloud computing and virtualization impacting way IT professionals integrate legacy systems with cloud-based ones and manage data centers, say industry insiders.
Written by Kevin Kwang, Contributor

Diminishing land space and increasing operating costs are driving the need for convergence in data centers. In turn, virtualization and cloud computing are helping to make this convergence possible but are also adding complexities, as IT professionals now need to know how to integrate existing physical infrastructure with virtualized ones and manage the environment efficiently.

Mayank Kapoor, research analyst of ICT Practice at Frost & Sullivan Asia-Pacific, identified decreasing datacenter space and increasing operational costs, particularly in countries such as Singapore and Hong Kong, as key drivers for convergence in the data center. This convergence, Kapoor added in his e-mail, is increasingly fueled by virtualization and cloud computing in general.

Adding to his observation, Cynthia Ho, associate market analyst of virtualization and data center at IDC Malaysia, noted that with virtualization and cloud computing, the "most significant impact" is the transformation of IT landscape from a traditional physical infrastructure to a virtualized one.

Elaborating, Ho said all systems, applications, security activities and IT services are now focused on locating and utilizing virtual resources within the data center instead of physical ones.

That said, she added that companies have not virtualized all of their hardware and there are still workloads in their data centers that run on traditional models. This, in turn, means the issue over how to integrate physical and virtual instances of compute resources, to effectively run one's IT shop, remains the main concern companies grapple with, she pointed out.

Kapoor also pointed out that while virtualization decreases the number of physical servers required to run the same workloads, these high-density servers also create new cooling challenges as more power is needed to prevent them from overheating.

Applications, too, are becoming "increasingly chatty" and with greater amounts of data being stored on the server, the traffic traversing the networks today has seen manifold increases, he added. This is why the demand for bandwidth and WAN (wide area network) optimization offerings has grown, said the Frost & Sullivan analyst.

Changing customer needs
Cloud vendors are also seeing changes with regard to what customers are asking for.

Caleb Ying, country head for managed services at Fujitsu Asia, for one, told ZDNet Asia in an e-mail that as customers consume compute resources in an on-demand, subscription-based manner, this implies that datacenter service quality will be determined by real-time customer satisfaction rather than contractual service-level agreements (SLAs).

"A glitch in the data center can potentially bring down the entire service for companies and, increasingly, service providers are looking at multiple data centers for backup," Ying explained. "Data centers not only have to be resilient, a single point of failure will not be acceptable, particularly for critical services."

Eric Goh, managing director of EMC Singapore, also pointed out that for most big enterprises, their data centers will consist of mainframes, different flavors of Unix systems with different chipsets and variants of Windows server operating systems. "There's little standardization, let alone automation in these environments," he noted.

"Customers have too many data centers and vendors, and have spent too much money on customizing applications, leaving the software rigid and inflexible and end-users with little control. These are points that customers want addressed," Goh said.

"[Companies are] willing to spend more on IT, but they're saying 'you've got to help us with this problem' [as] this [situation] is not sustainable."

Outsourcing datacenter operations
Quizzed if outsourcing data centers to third-party operators would help CIOs solve this conundrum, Steve Garrou, vice president of outsourcing and cloud services for Savvis, did not reply directly but pointed to how cloud computing allows IT to add more value and innovation, resulting in a "major impact" on companies' performances. Savvis in April was bought over by the third-biggest telco in the United States, CenturyLink, for US$2.5 billion.

"Many high-performing companies have one thing in common: they selectively outsource IT if it's perceived as either a high-velocity enabler of their business, or not a strategic function of their IT organization," Garrou explained. "By outsourcing these functions, companies can now focus on activities that are key to making their [core businesses] successful."

However, Kapoor noted that convergence and cloud computing alone do not represent a shift toward increasing reliance on third-party data center operators.

He explained that the last 5 to 7 years have seen an increasing outsourcing trend in the contact center and business process horizontals and this, in turn, impacts how businesses use IT resources as more organizations sign up for managed hosting and cloud services.

What is new, though, is how enterprises are separating their workloads according to those they are able to entrust to third-party vendors and what they need to keep in-house, he said.

"Certain workloads have moved into public clouds while some moved into private clouds," the Frost & Sullivan analyst noted. "Given that there are definite cost savings and efficiencies to be gained by engaging third-party service providers, the [cloud] market is expected to witness continued growth."

EMC's Goh added that private and public cloud models are "complementary", and the ideal situation for a large enterprise would be to "combine the best of both worlds"--access innovative and on-demand services hosted on public cloud, coupled with the management and tight control of an internal, private cloud.

In other words, he noted, companies should deploy a hybrid cloud model to deliver compute resources to users.

As for small and midsize businesses (SMBs), Kapoor noted that the requirements of these companies are different from large enterprises. SMBs' lack of capital and limited understanding of IT makes outsourcing to third-parties via cloud computing a viable option, he said.

"Given that SMBs are less concerned about security and privacy of their data, cloud provides them with the right value proposition to meet their IT demands," the analyst surmised. "Hence, they are not expected to build their own data centers in the foreseeable future.

Editorial standards