Savings, ROI main considerations for server refresh

Economic uncertainty is driving enterprises to emphasize cost savings and return on investment when doing a server refresh, said vendors. Virtualization and cloud computing are also affecting server purchase decisions, which in turn boost demand for blades, they added.

Ng Tian Beng, Dell's managing director for South Asia and Korea, noted in an e-mail that current economic and IT trends are influencing the procurement process and enterprises are employing greater caution when it comes to IT investments.

According to him, savings on Opex (operational expenditure) and expected ROI (return on investment) of implementing new technology are two key factors enterprises consider when doing a server refresh.

A spokesperson from IBM said improving the TCO (total cost of ownership) is a key consideration for Southeast Asia-based businesses when conducting server refresh exercises. Aside from initial hardware purchase, TCO takes into account datacenter space, power supply and consumption, flexibility to move workloads within a server or across different servers, performance stability, maintenance and support for service consolidation, he said.

Adding on, John Fruehe, director of server products for AMD Opteron, said in a phone interview with ZDNet Asia that enterprises looking to overhaul their server equipment are concerned with power efficiency and scalability. Server lifecycles, he noted, vary from company to company. Larger companies and government agencies refresh equipment every three years while smaller companies which are growing quickly will update their hardware every year-and-a-half to three years.

The need for power-efficient servers, explained Fruehe, stems from the expansion of data centers over the years which also leads to higher cost of power and cooling.

Companies are also looking at the scalability of their servers as they want resources that can easily grow with their business, he said, adding that this is especially true in Asia where businesses are developing rapidly.

A spokesperson from Intel agreed that performance and power efficiency are among the criteria for enterprises when choosing servers. In his e-mail, he added memory capacity and increased I/O (input-output) capabilities as well as open standards as some of the other factors businesses consider when refreshing their servers.

Elaborating, the spokesperson said: "Enterprises are turning to consolidation and virtualization to improve on server utilization rates. To enable this, servers with better performance, higher memory capacity and increased I/O capabilities are required."

As for open standards, he noted that for cloud computing to realize its full potential there is a need for data centers and cloud infrastructure to be more secure, efficient and open. "This will increase interoperability, flexibility and choice. Enterprises want the confidence to be able to invest in solutions today that will remain relevant into the future," he added.

Optimizing for virtualization and cloud
Dell's Ng concurred, noting the two technologies are "at the epicenter of the new decade of computing innovation". The executive claimed the hardware giant was first to market an embedded hypervisor as well as "fail-safe virtualization", which aims to eliminate single point of failure so that higher levels of availability can be provided to virtualized environments.

In terms of software, Dell has AIM (Advanced Infrastructure Manager), VIS (Virtual Integrated System) Self-Service Creator and VIS Director to help businesses manage their services and workloads while allowing end-user self-service provisioning and management of multivendor IT resources in a cloud environment.

IBM's representative pointed out that Big Blue has had "decades" worth of virtualization experience--since the late 1960s--with the first product released shortly after its early hardware such as mainframe machines.

Till today, IBM is "constantly looking at ways to provide comprehensive cloud infrastructure stacks [including] hardware, software and services, for private and public clouds, at all levels of the cloud journey", said the spokesperson.

In an e-mail to ZDNet Asia, Stephen Bovis, vice president and general manager for industry standard servers at Hewlett-Packard Asia-Pacific and Japan, pointed to an HP-commissioned research which found that nearly 90 percent of respondents consider their company's virtualization and cloud strategies to be related. The findings also revealed that 62 percent think that virtualization is the first step to a cloud strategy.

He added that the company's ProLiant servers allow applications to be moved "around anywhere, anytime in an extremely optimized way". "This improves IT flexibility and response to business requests, ultimately improving business speed and agility," said Bovis.

According to Intel, its current server processors--the Xeon 5600 series, Xeon E7 family and Itanium 9300--have virtualization technology built in. This provides benefits such as "acceleration of fundamental virtualization process throughout the platform", reduced latency and fewer potential bottlenecks, and improved short- and long-term "value for enterprises' software and server investments", he said.

Blade server demand continues
In an August report on the Asia-Pacific server market for the second quarter of 2011, Gartner said blade servers were the fastest growing category of servers. The research firm also noted that blades were the preferred platform for virtualization in large and midsize businesses.

HP's Bovis agreed that the growth and adoption of blade technologies will continue over the next year as customers shift their infrastructure to a virtual IT environment which will require new processes, infrastructure and management tools.

"Blade servers are an ideal foundation for enterprises planning to operate shared services environments or private cloud infrastructures. Businesses look to blade servers as a way to approach the server sprawl that has been driving [up] operational costs," he said. "By provisioning workloads onto compute-dense form factors, such as server blades, businesses can preserve datacenter floor space and contain power and cooling costs within the limits imposed by local power utilities."

Intel concurred, explaining that companies choose blade servers were picking up "primarily due to the higher density of servers per rack, thereby reducing footprint and datacenter space".

A Huawei spokesperson added that aside from a blade server's smaller form factor that can help reduce datacenter space, it provides better redundancy, fault tolerance support and better expansion support.

Blade systems also have a "hot-swappable feature" that allows customers to perform fixes to hardware without affecting other online servers--which helps to reduce downtime, said Huawei.