Lead

Data centres: tech on a platter

The process of consolidating distributed data centres to a central location has begun. Many companies are choosing to outsource that task while a slew of technologies is making life easier for data centre managers. Faiz Askari reports

With organisations becoming dependent on their systems and applications, there is a need to be able to access these applications from multiple locations. It is no surprise that it is necessary to co-locate servers within a data centre. “The data centre market in FY 2005-06 was around Rs 470 crore, and is expected to exhibit a CAGR of 38.1 percent,” says S Kaushal, Programme Manager, Information & Communication Technologies, Frost & Sullivan.

Every organisation has its unique data storage requirements. However, as companies the world over focus on core competencies and reduce costs, processes which were earlier considered core activities are increasingly being outsourced. A robust infrastructure can be extremely expensive and complex. Although businesses usually understand their own organisation and market better than any consultant, they often lack the internal expertise, dedicated resources and network access to plan and manage their own data infrastructure successfully. Hence this function is increasingly being outsourced to companies offering data centres and data warehousing services.

Third-party data centres

Today, companies across industry verticals are looking to outsource key parts of their business processes to data centres. Industries where extensive customer interaction is a given—automotive, telecom, banks, financial institutions—are early adopters. What’s more, even government and public sector organisations are realising the benefits of outsourcing key IT processes to data centre specialists.

Kaushal comments on the growing trend of data centres in India. “Enterprises in the country are increasingly outsourcing their data centre. The major driver for this trend is the fact that enterprises are faced with increasing pressure to optimise internal operations. A data centre helps in reducing capital expenditure, operational expenditure and total cost of ownership. Another factor that is driving the data centre is the desire of enterprises to concentrate on their core business areas.” Moreover, an expert service provider offers skills and resources which are not available internally; these include up-to-date technology and skilled human capital.

“In addition to this, since availability is the key, the redundancy offered by a data centre provider is often preferred in comparison to setting up one’s own,” states Ponnanna Uthappa, Head, Marketing, Team Computers. “Security is a major concern because important data resides on these servers. Data centre service providers are easily able to provide disaster recovery and business continuity since they operate from multiple locations.”

Architecturally, a next-generation data centre relies on certain pools of resources that can be combined to support a variety of applications. This architecture applies to the four critical aspects, and can be called the pillars of data centre infrastructure: storage, computing, management and networking. How can organisations transform their data centres to the next-generation model? The trick lies in translating this vision into a series of discrete, incremental steps—in other words, a road map. The road map comprises four major steps: consolidation, standardisation, virtualisation and utility.

With consolidation, multiple devices are consolidated into a single location. Standardisation ensures that devices have consistent interfaces and protocols. Virtualisation abstracts the physical infrastructure, creating one or more virtual (logical) instances running on a single physical resource. (For example, one physical server might be virtualised to appear as eight virtual servers, perhaps running different operating systems.) Utility describes an infrastructure that appears as a service for purchase on demand, similar to a utility such as water, electricity or phone service.

According to Soumitra Agarwal, Marketing Director, NetApp, “As data increases in terms of volume, organisations have acknowledged the importance of their data. In this case, where there is a need to store, maintain or archive data, people are thinking of getting this entire process outsourced to any data centre.”

Data is the focus of any data centre, and data storage, management and retrieval are critical disciplines. Data centre storage encompasses ‘live’ data, which is frequently accessed and processed, and various shades of ‘near-live’ data, which is stored on slower media or offline archival media. Key technologies are storage area network (SAN), network attached storage (NAS), virtual SAN (VSAN) and fibre channel.

Comments Praveen Cherian, Country Manager, Networking, Site and Security Services, IBM Global Technology Services India, “Over the last two-three years, organisations in India have started deploying blade servers in the data centre. With this, the power and cooling requirements have changed from 2 to 20 KW per IT rack. That means increased cooling distribution requirements inside the data centre, which has resulted in the adoption of new cooling technologies, one of which uses a raised floor as well as supplemental cooling for cold air distribution. Another uses a rack-based chilled water system without raising the floor. The focus is more on heat removal in this kind of design. Power and cooling designs have seen greater emphasis on redundancy and modularity. IBM has built over 200,000 sq ft of data centres for over 55 clients in India.”

Storage demands

Storage requirements are growing rapidly. The increasing use of electronic documents and processes is creating a glut of information for most organisations and a storage nightmare for IT organisations.

The types of data used by organisations continue to proliferate. The use of systems that contain structured, semi-structured and unstructured data types continues to grow, and they generate more business-critical information than ever before. This creates problems beyond just storage since this data must be indexed and must be easily retrievable at a moment’s notice from a variety of applications. Any system deployed within an organisation for data retention, extraction and destruction must support all these data types.

Talking about storage technology advancements in the data centre space, Sunil Sapra, Regional Director, Hitachi Data Systems India says, “Data centres are the result of a consolidation exercise for all or most applications to obtain better manageability, performance and ROI. In a data centre it is imperative that the entire IT infrastructure be aligned directly to the profile of business applications.”

Disaster recovery & business continuity

Disaster recovery and business continuity became buzzwords in the industry after the 9/11 attacks. In the recent past, natural calamities like the hurricanes in the US, the tsunami in Asia, and floods in Mumbai have dramatically impacted the thought-process of organisations worldwide and in India. Notes Manoj Chugh, President, India & SAARC, EMC Data Storage Systems, “For organisations across verticals—whether software, telecom, manufacturing, insurance, healthcare, finance, banking or government—domain disaster recovery and business continuity have become top priority. Organisations in India have to look at implementing back-up, recovery and archiving solutions that help them to store, archive, back-up and restore their critical data and applications. Additionally, they also need to look at business continuity solutions to ensure that while restoration of site, data and applications is going on at the disaster site, the business continues to operate from another site and customers benefit from 24x7x365 application availability. In such a scenario, the concept of data centre is getting a lot of importance since all these demands of customers can be met by, and outsourced to, a data centre.”

Networking the data centre

In this day and age, data centres are characterised foremost by the size of their operations. A financially viable data centre could contain anything from a hundred to several thousand servers. This would require a cohesive network architecture that supports immediate data centre demands such as consolidation, business continuity and security.

Sumit Mukhija, Business Development Manager, Cisco Systems India & SAARC says that “it is critical for a data centre to ensure maximum data security and 100 percent availability. Data centres have to be protected against intruders by controlling access to the facility, and using video surveillance. They should have the capability to withstand calamities like fire and power failures. Recovery sites have to be maintained where everything in the data centre is replicated. Data centres provide a shared, multi-host, multi-application environment to carry out the hosting of large volumes of corporate data along with providing functionalities like data mining and data warehousing. As businesses go global and get Internet-enabled, these services become mission-critical. There is a requirement for services from simple Web-hosting to managed services such as storage on demand, performance measurement and storage management.”

A data centre network architecture should comprise three layers:

• Foundation Infrastructure: Including the intelligent IP network infrastructure, intelligent storage networking, and data centre interconnect
• Network System Intelligence: Including security, delivery optimisation, manageability and availability
• Embedded Application and Storage Services: Including storage virtualisation, data replication and distribution, and advanced application services.

Security demands

For security purposes, most organisations have started using biometric-based access control for critical areas inside the data centre. IBM’s Cherian adds, “Data centre have seen increased use of integrated building management systems which have capabilities of remote monitoring using IP addressing. Integration of data centre systems with network management systems has also empowered IT managers to manage their data centre more effectively. With increased complexity in data centre design, system integration and project management play a key role in managing time-lines, scope and budget.”

It is essential that security and network managers collaborate to understand the particular vulnerabilities and threats to data centre resources so that they can develop a robust network security architecture. Opines Mukhija: “Vulnerabilities and threats can prevent users from accessing mission-critical applications, directly disrupt application operation, or compromise confidential and valuable information. Enterprises need a data centre architecture that offers IT and network managers an end-to-end, defence-in-depth security strategy and solutions to prevent or contain data centre attacks.” On another technical aspect of security at the data centre, Jasjit Sawhney, the CEO of Net4 says, “One of the prime purposes of a data centre is to keep the data and applications secure. The organisation takes into consideration the robustness of the technology used to keep the data safe.”

Emerging trends

“Companies are installing blade servers to address space constraints. Although blade servers occupy less space as compared to traditional rack servers, the heat output of blades is higher. These companies will have to invest in cooling systems to address the higher heat output of these servers,” points out Kaushal of Frost & Sullivan.

Grid computing has several intersections with utility computing, and shares these attributes: dynamic resource discovery, automated resource allocation, workload scheduling, and coordinated data migration. Grid computing has evolved out of high performance computing (HPC), which has high dependencies on storage including capacity and performance. States Chugh of EMC, “Many commercial data centres have supported HPC solutions for various departments in their company depending on the industry. For example, automotive and aerospace industries have a lot of mechanical CAD analysis, life sciences have DNA and compound analysis, financial services have portfolio and risk management, and so on. Government HPC applications include weather, satellite feeds, and many areas of advanced research. To augment the compute-centric side of grid computing, some areas of HPC research are keeping terabytes (and years) of research data online, and building portals and search engine front-ends to serve researchers.”

Virtualisation within the network further aggregates pools of storage across several storage arrays, enabling a number of features including reduction of planned down-time for data movement across arrays, volume allocation to meet service level agreements in a tiered storage environment, and data movement in support of an ILM strategy. Concludes Chugh: “EMC believes that virtualisation in the network should be out-of-band to enable scalability of the environment without constraint of the I/O path.”