Following on from my first blog post “Cloud Data Centers and the Art of Feng Shui”, I thought it might be interesting to look at four principles which are guiding the way that Cloud Data Centers are being constructed in China.
As in any other region in the world, Cloud Data Centers in China must become operational in a very short time window, as well as be able to scale according to user demand for services. At the same time they must be cost-effective both from a construction and from an operational point of view. By meeting these criteria, the data center is able to meet the business need, ensure end-user satisfaction, and meet the financial requirements for a robust business model. Four principles can be applied to ensure that data centers meet these goals:
1. Standardization of Data Center Infrastructure. Standardization can ensure that customers are able to obtain very high levels of data center performance at the same time as achieving very rapid facility deployment. With legacy data centers, the complete cycle from initial design through to the operational phase of a new facility can take two or three years. Through standardization, the cycle can be compressed to as short as eight to nine months. Standardized, data center infrastructure components are pre-configured, preset and pre-tested at the factory, so that deployment time can even be shortened to a few weeks.
2. Increase the Density of Data Center Equipment. Moore’s Law continues to affect us all with chip performance increasing almost exponentially, and the cost of performance dropping almost as quickly. The effect on the data center is profound because facilities built for a 10-year plus lifecycle need to be adaptable to the needs of technology which increases in density every 18 months, and sometimes less.
The answer is to improve the IT rack density. In China, there are a lot of data centers that occupy a very large area, but at only 3 or 5 kilowatts per rack position, their density is not sufficient to support a rapidly evolving IT load. Our research shows that when data center density is increased dramatically, for example to 20 kilowatts per rack position, there are not only energy efficiency gains to be made, but also a significant reduction of facilities (or, grey) space.
3. Take a Modular Approach. Modularity enables data centers to be scaled according to demand, taking a pay-as-you-grow approach. In other words, data center operators can purchase the appropriate infrastructure based on business growth. Implementation and deployment can also take a modular step-by-step approach. In this way, you not only achieve the energy saving effect of right-sized infrastructure, but also conserve capital budgets. Moreover, the modular approach also endows the data center owner with adaptability and agility, e.g., the containerized data center can be located anywhere in the world.
4. Integrate management software. Schneider Electric has been establishing a track record for developing critical power and cooling equipment together with integrated IT management tools. Recently, Schneider Electric introduced StruxureWare for Data Centers DCIM software suite, which can manage the building, facilities and infrastructure. DCIM can also communicate with enterprise and IT software from major vendors including Cisco, Microsoft and VMware to provide interconnectivity and interactivity. Software has a major impact in the way that current and next-generation data centers are being architected and managed for maximum availability and efficiency – two imperatives which are vital to success for Cloud Data Centers.
We’re applying these four principles with great success in China, but do you think that there are others which could be considered? Or do you think that some of these principles are region-specific and not applicable to the challenges of your market? If so, leave a comment – I’d like to learn about other experiences from around the globe.