Topics in this article
With the ability to scale quickly without linear add-on costs, a hyperscale data center will change every aspect to the data center
In the digital age, businesses need to take decisions on a real-time basis. From booking cabs (Uber, Ola) to consuming food (Swiggy, Zomato), consumers are increasingly shifting their preferences to online mediums. Given the extreme pace that businesses operate today, organizations cannot wait for their IT systems to deliver the services at a pace that traditional systems offer. They need an IT infrastructure that can scale quickly to provision increased demand and scale down appropriately when demand reduces.
This has led to the demand for hyperscale data centers. Hyperscale refers to the capability of an IT system or architecture to scale exponentially and rapidly to respond to demand that is increasingly heavy. Given the demands of the digital era, hyperscale data centers are set to dominate data center trends and design. A report by Allied Market Research titled, 'Hyperscale Datacenter Market by Type', states that the global hyperscale data center market is expected to generate USD 71.2 billion by 2022 from USD 26.33 billion in 2017, growing at a CAGR of 20.7% from 2017 to 2022. Another report by Markets & Markets estimates the hyperscale data center market to grow from USD 25.08 billion in 2017 to USD 80.65 billion by 2022, at a CAGR of 26.32%. Cisco estimates that by 2021, traffic within hyperscale data centers will quadruple, and hyperscale data centers will account for 55% of all data center traffic by 2021.
Let’s look at some major ways that hyperscale data centers are reshaping IT, by encouraging data center architects to consider some vital parameters:
- Modularity: Thanks to the Lego-like approach, enterprises can take a step-by-step approach to planning for their data center. This means that they can replace individual physical components and design their data center. This approach helps organizations in adapting themselves progressively to a design that is far more agile and flexible to their current short term needs, while staying on the long-term path.
- Redundancy: Due to the modularity, data center operators can replace individual physical components rather than the traditional approach of replacing the entire server, which not only increases costs, but also increases the downtime. This approach also gives extreme flexibility in scaling at the physical level, as components can be added modularly. For example, in case a server fails, an application can be moved from one server to another server, without downtime. This helps in maintaining a high level of system availability as there is no single point of failure. This approach also helps in analyzing different points of failure, especially when there are many components that could have caused the issue.
- Reducing cost of downtime: In the digital age, the cost of downtime can be huge. With automation, this can be reduced as most of data center outages are caused by human errors. For example, in the highly publicized British Airways incident which led to more than 75,000 passengers being stranded in one day, this was caused by human error. An engineer had disconnected the power supply source at a data center near the airport, and when it was connected back, it led to a surge of power that caused the IT systems to crash. As most processes are automated in a hyperscale data center, errors like these could have been avoided.
- Energy savings: From using AI in the data center to reduce energy costs, to using new energy cooling techniques such as liquid cooling, hyperscale data centers are setting the standard for saving power.
As we can see, hyperscale data centers are clearly defining the future landscape of enterprise IT. With the ability to scale quickly without linear add-on costs, a hyperscale data center will change every aspect to the data center — from the way they’re designed, to the way hardware components are sourced, to defining the standards in terms of energy efficiency. These clearly are exciting times for the data center industry as we embark on a new dawn of enterprise IT where everything is software defined!