Archived Content

The following content is from an older version of this website, and may not display correctly.

The demands on data center infrastructure and space are having a profound impact on the industry today. Gone are the long data center build times as CFOs increasingly demand faster responses to the problems being experienced in the modern data center.

Insatiable appetites for data, rapidly aging infrastructure and redundant technology all contribute to the desire to roll out new applications and platforms, and all this with speed but surety in the maturing age of the Cloud.

Commercial Property Management group GVA’s new director for national markets, Mark Larard, says speed-to-market is the most current driver for data center developers. “Those few who can meet the timescales and the locational factors will inevitably see success.”

So it is no surprise that a new buzz term is being bandied about the data center industry.  It is called the ‘Just-in-Time’ data center, and the idea is it can save operations in a time of crisis. Interestingly, this idea seems to play across a number of industry areas – from the modular and containerized data center market to the enterprise equipment vendor, cloud provider and, getting back to GVA, the real estate manager.

Just-in-Time database
Oracle’s chief technology architect John Abel says he has seen the need for what he calls Just-in-Time data centers expand from the financial services and telecommunications sectors. Here the industries have already consolidated more than most but are grappling with rising amounts of data, and the need for growing amounts of compute power. “They can no longer sustain the growth that is required,” Abel says. “The result is that we are now getting into an area with customers that want to have physical project times of four-five months, instead of one-two years.”

Abel says that more customers are looking for storage compute resources with everything in one box. This is the sweet spot at which Oracle is aiming its Exadata database appliance and SPARC server family. It is also where it uses the term ‘Just in Time’.

For Oracle, Just in Time works in a box. “Exadata provides an engineered system with storage compute resources, servers, networking and storage all in one box,” Abel says. “SPARC allows them to plug in their original storage, and have compatibility with Solaris 10 and a smooth journey with Exadata, which makes the business case more compelling.”

The financial markets are moving to these engineered systems because of the short timeframe it can take to set up such dense compute systems. “It is just in time – it is inside the boundary of their projects,” Abel says.

“These companies only really have three options to get growth. They can re-engineer the application to make it scale, buy expensive compute resources and hope their solution can be elastic, or find an engineering platform.”

The development of the Cloud is further propelling this notion. “Many customers are doing R&D in either a private or public cloud. They are also looking for a route to fast acceleration,” Abel says. In this case, they will quite often rely on a cloud provider to test solutions before bringing applications back in house and purchasing equipment such as that outlined by Abel for fast deployment.

Clouds — Just in Time  
Virgin Media Business formed a cloud computing partnership with Savvis last year which also relies on companies needing compute in the nick of time. Telecommunications and financial services companies might be looking for in-house solutions, but mid-sized corporate customers and the public sector driven by cloud initiatives are more likely to outsource their Just-in-Time needs.

“We are seeing a lot of companies running or owning their own data center or hardware infrastructure that have had aging estates and need hardware upgrades. They want to go back to focusing on their core business,” says Andy Perrin, Virgin Media Business senior product manager for cloud services.

“In most cases, the CTO or CIO has been tasked with making the IT department more agile and flexible, and this leads them to look at adopting new solutions.

“We also see customers that are continually bringing out new products. We have one customer in the financial services sector that just wants to scale up and down as required, and when they don’t know if a new project will be successful, they prefer not to buy a bunch of IT equipment upfront. They will instead set up a virtual data center with us.”

This Just-in-Time concept – which is initially pay-as-you-go – works well for the provider, says Larard. Often, when the time comes to bring solutions back into the corporate data center, “it rarely comes back in house”.

“There are two factors behind this. The data storage inevitably grows, and as soon as space is freed up though strategic planning, the day-to-day data will grow to fill the newly created void. Outsourcing also allows the third-party provider to demonstrate capabilities and cost-effectiveness,” Larard says.

Just-in-Time container
For some companies, as seen with Oracle’s example, outsourcing to meet tight deadlines is simply not an answer.

HP offers a perfect case in point with the marketing of its HP PODs (performance optimized data centers) – containerized data centers that can be delivered in a short time frame to support all IT needs. HP’s most recent POD, the EcoPOD, even comes with its own power and cooling.

HP EMEA business manager Bertjan de Herder expects HP’s PODs to sell well in the Middle East and Africa, where infrastructure is required in a short timeframe to meet
the demands of a relatively new data center market. The POD’s main selling point is its delivery timeframe.

When Airbus required a supercomputing facility in Toulouse, France, last year, HP’s POD was used due to its high density and fast delivery time (more on that in the next issue).

HP head of Enterprise, Servers, Storage and Networking in EMEA Peter Ryan says PODs are attracting interest from clients facing difficulties in securing enough power to build out and time to fix their current space.

“You can take a POD to where the power is – you can link it to your existing facilities, you can have it pre-packaged, pre-configured and dropped into the data center space,” Ryan says. “For companies such as Airbus, to get access to as much compute power and storage power as they needed in the time they need it is an incredibly efficient way.

“There is a lot of legacy stuff built up in data centers over the years, and the amount of money and time they are spending on it is preventing them from accessing more modern technology that can deliver a lot more benefits.”
 

This article first appeared in DatacenterDynamics FOCUS magazine, Issuue 22. To receive copies of focus and sign up to our digital editions, click here.