Power Conditioning

Evolution not Revolution

Designing and managing the power infrastructure is one of the most crucial elements to be considered when designing a data center. Subodh Tagare, Marketing Manager, APC India, details the design parameters to be kept in mind and provides best practices to manage the data center power infrastructure. Anil Patrick R

What are the factors to be kept in mind when designing the power infrastructure for a corporate data center?

Despite revolutionary changes in IT technology and products over the past decades, the design of power infrastructure for mission critical installations like data centers and network rooms has changed very little. Although IT equipment has always required electrical power, the way IT systems are deployed today has created new power-related problems. These are of great concern to enterprises since most of these were not foreseen when the powering principles for the modern data center developed.

The physical infrastructure is the most important aspect that has to be kept in mind while designing a data center. This is mainly because this forms the foundation for the IT network to function effectively. One is the physical infrastructure or Network Critical Physical Infrastructure (NCPI) and the IT equipment which the infrastructure powers.

The factors that need to be kept in mind while designing the power infrastructure for corporate data centers are many. First of all, the systems should be modular in nature and capable of growing with increasing requirements. Secondly, these have to be pre-engineered solutions that eliminate and/or simplify most planning and engineering. The next factor to keep in mind during design is the modular and portable power infrastructure that can be moved to other facilities when required. Power metering is also a crucial component to be kept in mind during design.

What are the impediments faced by CTO/CIOs during set up of a data center power supply framework?

During a study carried over 30 months, APC focused on NCPI for server rooms/data centers and found a long list of issues faced. These issues were in the areas of availability, lifecycle costs, scalability, manageability and serviceability. When it comes to the issues specific to CIOs, CTOs (CXOs), the three biggest issues were driven by Business Value or Return On Investment (ROI):

Availability Issues: IT applications have become the life blood of many organizations and any non-availability can prove to be disastrous.

CXOs are concerned about:

• Single points of failure in NCPI e.g. a single generator or single aircon
• Long Mean Time To Recovery (MTTR) for NCPI. While IT equipment has become hot swappable and easily serviceable, NCPI remains susceptible to long technician repair times, quality and experience of the technician etc.

Agility: Today IT is seen as providing an edge to organizations. A quick deployment, ability to scale, portability etc. are the long required needs for IT deployment. However, these are bottlenecks for current NCPI architecture and this is a big challenge for CXOs.

Total Cost of Ownership (TCO): Due to the rigid nature of NCPI, enterprises are forced to deploy data center NCPI upfront and make many guesses on future growth. Due to the high risks involved if an organization undersizes NCPI, there is a tendency to oversize NCPI. This drives TCO of NCPI deployment to very high levels in today's data centers. As per our study, on an average, 70 percent of NCPI never ever gets utilized and is effectively wasted. This is definitely something that CXOs will like to do away with.

What are the usual mistakes made by organizations when they implement the power supply infrastructure for data centers?

NCPI is the foundation upon which the organization's IT is deployed. An ineffective NCPI deployment can result in NCPI availability being affected. These downtimes can occur due to equipment failures and poor architectural designs leading to redundancies in the wrong places. Hence, this translates to inadequate protection from failures or long Mean Time to Recovery.

Any NCPI issues have a potential to effect the IT setup of an organization and hence the whole organization. For example, NCPI setup in a data center not being deployed correctly could result in the cooling getting effected when power goes down. This can lead to IT equipment getting overheated and eventually a need for shutdown or face a failure of IT equipment.

Can you suggest best practices to ensure that enterprises get the best value out of their power investment?

Network-Critical Physical Infrastructure (NCPI) is the foundation upon which the networks reside. This foundation consists of power, power distribution, racks, cabling, cable distribution, cooling, cooling distribution, security—as well as the service and management for these applications.

In order to maintain a highly available data, communications, or converged network, four main ingredients are required:

• People
• Process
• Technology
• Network-Critical Physical Infrastructure (NCPI)

These four layers are completely dependent on each other. If even a single ingredient is missing, the business process(es) that are dependent on the network become(s) unavailable. Therefore, the layers represent a chain whose availability is only as strong as its weakest link.

Anil Patrick R can be reached at

anilpatrick@networkmagazineindia.com

Storage Technology

Utility computing is compelling

Storage major EMC believes forging partnerships and leveraging on combined strengths to address the sticky issues with Storage, and penetrate segments such as SME. With plenty of alliances on the anvil, Tony Leung, Managing Director-Marketing, Asia Pacific/Japan, EMC Corporation updates Venkatesh Ganesh on EMCs new strategies and products

SMEs shy away from the complexities of implementing networked storage. How have you tackled this issue while pursuing the SME segment?

We have announced our partnership with Dell in distributing the CLARiion AX 100 array. We understand that modern day storage solutions can sometimes get complex. The CLARiion AX 100 precisely aims to reduce the complexity of implementing network storage. The Dell/EMC AX 100 array is available in a direct-attached SAN-ready configuration. We also have a complete scalable SAN configuration too.

Our understanding in this space has been that the availability and performance benefits of networked storage are of utmost importance to the SMEs. The complexity and high cost previously associated with the technology, which deters small companies, is addressed by the CLARiion AX 100. We took into account these issues and decided to offer a cost-effective, and easy-to-use solution. This can enable customers to move from internal server-based disks to external direct-attached storage (for both SAN and NAS environments).

What features in the new product are most suited for the SMEs?

Firstly, the product (CLARiion AX 100) is based on an EMC patent-pending design and technology. It is a customer-installable networked storage system that is capable of storing up to 3 terabytes of information. Besides, there are built-in functions that are easier to manage and offer good protection. It incorporates the RAID technologies from the CLARiion CX family and delivers cost-effective ATA storage in a 3.5-inch rack-mountable enclosure with integrated storage management. The storage management functions range from monitoring local space-saving snapshot replication (for instant backup and recovery) to a user-friendly interface. The CLARiion AX 100 offers a web-based interface for secure and remote management of the system from any Internet connection.

SANs seem to be getting more complex. How is EMC addressing the issue?

Yes, we do agree that SANs are getting more complex. To address this need, we are developing a storage router software that will use virtualization technology. Basically, it will let customers migrate data among the disk arrays in a storage area network without shutting down or affecting the applications using the data. In the initial stages, we are planning to tailor the storage router for customers in large enterprises who deal with bigger and complex SANs. As this growth occurs, the tolerance for planned and unexpected downtime of storage arrays shrinks.

We have partnered with switch vendor Brocade Communications Systems and are looking at partnering with Cisco too (at the time of going to press sources within EMC are on the verge of announcing a tie-up with Cisco). With this partnership, we plan to develop a technology that builds intelligent storage routing capability into the networking companies' switch fabrics. The result will allow dynamic routing of data. EMC will work with standards organizations to figure out the best way to provide APIs, so that a storage router can work with intelligent switch fabrics. Besides, this virtualization will extend across all vendors.

Where does the concept of utility computing stand today?

We feel that the concept of 'utility computing' is compelling to say the least. From the CIO perspective, they are trying to do a couple of things. Firstly, they are trying to figure out whether it is hype or reality. Secondly, does it apply for my business? We feel that unlike the past 'big things' in technology, this wave of innovation would not require corporations to rip out existing technology and replace it with newer and costlier software and hardware.

Instead, they can gradually add technologies or services that make their computing systems more automated. It is about a rip-and-replace approach, but one that can incrementally move down the path of utility computing.

As a result, much of the cost and complexity can be wrung out. The idea is that the power plant-like computing systems of the future will operate both at remote data centers and within a company's offices. The idea of making information technology as easy as plugging into an electrical socket would be the key differentiating factor of this trend. Besides, the practical implications of aligning business requirements with IT and driving operational costs down would make this trend happen.

We feel that more than the hype surrounding utility computing, the key to explore this trend would lie in its interpretation for everybody.

If you take an example of storage management, there are so many messages that emanate from different vendors that it gets the CIO confused in trying to figure out the right solution for their businesses. Even the CIOs understand that sometimes the vendors' strategies are articulated only to apply to their environments.

Customers can look at storage companies to be their de facto storage management layer and weave it around the utility computing layer. This would simplify the existing complexity to a large extent.

What products are you launching in India?

For the Indian market, we recently announced a slew of products totaling more than twelve. Along with this, we have innovations on the software front, which extends across the entire range of EMC tiered storage platforms. These new software offerings are intended to enhance the backup, protection and recovery of information.

These products are targeted at the high-end and mid-tier SAN, NAS and CAS (Content Addressed Storage) segments. For SAN, we have CLARiion CX series and Symmetrix DMX-2 systems. In the NAS space, we have introduced the Celerra NAS gateways and integrated systems and to address the requirements in the content addressed storage segment, we have Centera, a system that unlocks the value of fixed-content digital assets stored in mainframe environments.

In addition to this, EMC last year announced a five year multi-million dollar investment and are backing it up with the industry's comprehensive and technically advanced storage software, platforms and services. We can meet the needs of the mid-tier storage market, which is one of the fastest growing segments in India.

Identifying this need, we have established partnerships with Dell and Datacraft (who are our global patners too) and an Indian network of system integrators like Wipro, Tata Elxsi and HCL.

The writer can be reached at

venkatesh@expresscomputeronline.com