Organisations make use of data for their day-to-day working and operations. Data facilitate the decision making process. Data too helps in identifying patterns for enhancing the services and analysing the weak links.

Organisations deal with data either by establishing data centres or by adopting clouds. Some organisations combine these two methods to use a hybrid model.

Data Centres

A data centre is a physical facility or space of networked computers and its associated components, say telecom and data storage to organise large amount of data. It then processes, stores and disseminates the data.

Small and Mini Data Centres

These data centres are set up on-premises or close to the users in business districts. The advantage is reduced response time, often measured in milliseconds. There is the lowest level of latency to meet real-time data computing demands. These are micro and mobile data centres which are suitable for use in applications, e.g. instant data centres, remote office and branch office, and edge computing. They are high performing and energy efficient.

A small data centre can be as small as a few racks of servers or about the size of a small room. These are flexible, and can be scaled up. They are suited for densely populated markets. Improved connectivity will enable small data centres to serve local business needs at competitive rates.

Large Data Centres

These are gigantic facilities with thousands of servers processing so many terrabytes of global data.

The biggest data centres are in China. Their size could be as large as 7-10 million square feet. They consume humongous amount of electric power. The size of the data centres in Europe and the US too could be 3-7 million square feet. Most such data centres are located in remote areas away from the urban areas.

These big centres offer economies of scale. The data is expanding exponentially. This requires large data centres. But all regions and countries cannot afford such a set up.

Demand for Data Centres

The highest demand for data centres is in the Asia Pacific region. China and India are the leading players here. Corporates such as Oracle, IBM, MS, Google and AWS are investing in data centres in India.

In India, the roll-out of 5G imninent. The atmosphere is conducive for data centres. RailTel proposes to set up 100 mini data centres across the country, mostly in 2-tier/3 tier towns.

RailTel is expanding its optical fibre network laid along the railway lines. The edge computing data centres will be linked to the optic fibre network. Big, small and medium size data centres will facilitate the growth of digital India.

Types of Data Centres

Telecom data centres are operated by telecoms. Enterprise data centres are built and owned by a company, either onsite or offsite. Co-location data-centres provides cooling to multiple enterprises within one data centre to hyperscale the customers. Hyperscale data centres are owned and operated by the company itself.

Cloud

Either individual or collective, a group of services is provided by a network of services which possess a unique function. A could is not a physical entity. It is a group or network of remote servers arched together to operate as a single entity for an assigned task.

A cloud is built with a lot of computer systems. A cloud is accessed through internet. Cloud providers offer cloud as a service. It lets the user rent the computer systems.

Cloud compute is on-demand delivery of IT resources over the internet with pay-as-you-go pricing.

Cloud also helps the storage of data. In cloud services, a company need not own or maintain physical data centres and servers. Still, the organisation can access technology services — computing power, storage and data bases on as-needed basis from a cloud provider.

Amazon Web Services (AWS) is a cloud provider.

Types of clouds

Public cloud is open to all on pay-as-you-use basis. Private cloud can be accessed with the permission of the organisation. Hybrid cloud methodology is a combination of public and private clouds, and serves different needs. Community cloud is a methodology to offer services to a group of people in an organisation or a single community.

Difference between Cloud and Data Centre

• Cloud is a virtual resource. Data Centre is a physical resource.
• With less investment, a cloud is scalable. A data centre requires more investment for scaling up.
• A cloud attracts less maintenance costs, as the service provider maintains it. A data centre has high maintenance costs.
• In cloud the service is provided by a third party. In data centres, we have our own developers.
• Cloud is high performance entity as compared to investment. A data centre is low performance entity as compared to investment.
• One must develop a plan to customize a cloud. A data centre is easily customizable, without any hard plan.
• An internet connection is a must for cloud. A data centre may or may not have internet connection.
• A cloud is easy to operate and viable. A data centre requires experienced developers and may not be viable.

Additional Comments

Modern data centres have software defined networking (SDN) to manage traffic flows via software. Historically, organisations used on premises data centres. The necessary infrastructure is maintained on-site. There are services, support web, email, networking hardware and uninterruptive power supply (UPS).

Organisation are moving away from onpremises data centres to cloud data centres. These are virtual data centres which can be scaled up or downed by a few mouse clicks. They offer Infrastructure-as-a-service (IaaS) and Platform-as-a-service (PaaS). Clouds too have issues of cost, lack of viability, accountability and transparency. Cloud provider is responsible for maintenance, updates and meeting service level agreements (SLAs).

All this does not mean moving everything to cloud. There are hybrid cloud data centres — a mix of on premises data centre components and virtual data centre components. These are shared responsibility models.

Data management is done across cloud, core and edge. A lot of infra remains on-prem(on-premises). However, there is storage complexity. Banks no longer have tiered storage, and data is to be available in real time. Previously, data used daily was put on faster discs, and data at rest on slower discs. These days even historical records are needed for analytics. We require a unified data operations approach.

We make use of descriptive analytics, diagnostic analytics, predictive and prescriptive analytics.