Trends in Data Center Optimization
Vaibhav Bhatia is a Senior Consultant with the Sustainability Practice at Infosys Limited and a certified Data Center Associate. With 9 years of IT experience, he has managed operations of a data center, and worked on several data center optimizing and Green IT initiatives.VAIBHAV BHATIA
In Part 1 of this article, we highlighted the pros and cons of various data center designs. In this the concluding part, we look at some of the trends that have a significant impact on data center architecture.
Energy cost continues to be the fastest growing data center expense and is reportedly exceeding the cost of computing equipment itself. Power Usage Effectiveness (PUE) and Carbon Usage Effectiveness (CUE) are internationally recognized metrics, with more metrics being defined which will become industry standards. New data center construction and modification will have to jointly cater to the business and the planet.
With costs spiraling in the data center, focus is turning towards optimizing the data center infrastructure. There are several emerging trends which demand notice and going to be on the decision table during the coming year:
(a) Remote Monitoring for the Data Center
The argument to outsource or not continues. Companies that have decided to keep their IT in-house are finding it cost effective to use external data center monitoring providers. External monitoring and first-level support requires secured access to IT and, in some cases, physical infrastructure devices. The need for additional IT infrastructure firewalls and security will add to the complexity of the data center. On the other hand, personnel and dedicated work space are continuous expenditures. Also, scaling the team size is easier in an outsourced environment.
(b) Data Center Site Location
Selecting a data center location is an important decision. Because technology now allows for most tasks to be performed remotely and only skeletal staff is required onsite, a wide range of geographies are available for sites. Some factors influencing this decision are:
- Climate-neutral locations
- Low expense per Kilowatt hour
- Minimal environmental hazards
- Lower cost of living
- Less densely populated areas
- Low construction costs
- Low labor costs
- Feasibility of “free cooling” – where air temperatures from the outside are used to cool the data center, requiring less mechanical cooling.
(c) Green IT
Energy consumption directly impacts cooling expenses due to heat from devices. Strategic business plans will directly impact the type and number of IT devices installed in the data center. Knowledge of device types and their efficiency is important as it will impact the power and cooling strategy of the data center and the physical design of the data center. Making IT more efficient and cost effective continues to be an industry focus, both the manufacturers and the consumers.
(d) Scalability & Modularity
In years past, only hardware and software needed to be scalable and modular in architecture to cater to growing demands. Cost and demand pressures on the infrastructure now necessitate a scalable and modular design approach for the data center infrastructure.
Such an approach, for example, applied to the UPS and Power Distribution system would allow one to add/disable sections of the data center without affecting another. Flexible designs enable hosting providers to add and remove sections of the data center based on the customer’s requirements. Catering to customer’s differing loads and mandates are also possible in a flexible design.
(e) DR Optimization and Re-usability
Historically, hardware utilization is low and virtualization has come a long way in addressing this. However, the focus has always been to improve the efficiency of the production data center, and the DR/backup center is left out of consideration since it can be “turned off” or “idle.” Organizations now realize that a significant investment has been made in these “idle” data centers and are coming up with innovative methods of using the Disaster Recovery (DR) infrastructure. The trend of using the DR center for testing purposes, training and staging will continue to grow. Designs will need to take into account the ability to switch to production in the safest and quickest means possible.
(f) Data Center Infrastructure Management (DCIM)
The convergence of IT and Data Center facilities management is a reality. Most facility infrastructure devices are IP addressable just as IT devices have been for years. Development is underway by several companies to marry the two former disparate environments. The capability to conduct “what-if” scenarios for planning IT components and physical infrastructure elements will enable designers to know how plans and changes in the IT environment impact the physical environment and vice-versa. For instance, what will the load be on our UPS and cooling systems if we install a specific type of IT hardware? Cost avoidance by not over engineering will be realized as these products penetrate the market.
It is imperative that these considerations be taken into account at the early stages of data center strategy and design as these features can make a significant dent in IT budgets and can have a detrimental effect on capacity and cost if not planned well.
Please note: Ron Diersen, Lead Consultant with the Sustainability Practice at Infosys Limited, with more 30 years IT experience, 26 of them within the data center, also contributed to this article. He has held various responsibilities in the construction of new data centers and the build out of existing sites, working on both physical and IT infrastructures.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.