Data gravity is useful planning tool for Digital Realty

Five factors influence data gravity: enterprise data stewardship, mergers and acquisitions, digital-enabled interactions, data localisation and cyber-physical aspects.

Digital Realty is well-positioned to serve customers with a strategy that is based on "data gravity", a concept coined in 2010 by Dave McCrory, now VP of Growth and global head of insights and analytics at Digital Realty.

Data gravity refers to the propensity for bodies of data to draw applications and services closer, and could affect an enterprise’s ability to innovate, secure customer experiences, and even deliver financial results on a global scale, Digital Realty said. The company has developed the Data Gravity Index*, which measures the creation, aggregation and private exchange of enterprise data and examines its impact on the Forbes Global 2000.

The concept studies what happens to data and the implications of this. For example, data has to be generated, stored, moved and analysed. "It isn't only the data coming in, it comes in, it sits and then it grows," said McCrory. The analysis leads to a feedback loop, "causing a large amount of growth in both demand for being able to store processed data, but also where all this equipment must be housed and what do you do with all that data," he explained.

At the same time, there are currently many trends accelerating data creation lifecycle, so the amount of data available is growing quickly. The five main drivers are data stewardship from enterprises, mergers and acquisitions, digital-enabled interactions, data localisation**, as well as the cyber-physical integration of security systems**, McCrory said.

McCrory came up with a formula for data gravity, basically data mass x data activity x bandwidth, divided by latency squared. A larger number is better than a smaller one.

- Data mass refers to data at rest, data that's being stored, the actual quantity of data;

- Data activity is data in motion, moving across the network. It occurs when data at rest is being accessed;

- Bandwidth is defined as the bandwidth measured for a specific metro area. McCrory used the analogy of the number of lanes on a highway. "The more lanes means more traffic and flow which would be an advantage," he said.

- Latency is analogous to the speed limit on the highway, and would be a measure of time.

"Latency is the great inhibitor," McCrory said. "The greater the latency, the less the gravitational effect."

According to McCrory the worst case scenario would be waiting so long for the data to arrive (high latency) that it would basically be considered downtime, driving people to look for a different system to use. On the other hand, a low latency of - as an example - 1 millisecond would be quite acceptable, and would not impact anything.

With the numbers, McCrory said a data centre strategy cannot be application- or backhaul-centric. "It's no longer a viable architecture decision to believe that the data can always be fit to the processing," he said. "The future is bringing the processing to the data."

One less than ideal scenario could be a slowdown for edge devices if they have to wait for data to come in from a few centralised locations the cloud, McCrory said. It would be quicker to have the data shared in more locations, such as at colocation facilities, so that it can be accessed more quickly.

In September 2020, Digital Realty shared that:

- Data gravity growth is expected to double annually through 2024 as data stewardship drives global enterprises to expand their digital infrastructure capacity to aggregate, store and manage the majority of the world’s data. Where the data is located will become much more important to global enterprises so that they can meet compliance requirements by maintaining local copies of critical data.

- Enterprises are approaching quantum computing levels of data creation, processing and storage. The Forbes Global 2000 enterprises across the 21 metros analysed are projected to create data at a rate of 1.1 million gigabytes per second by 2024, will be required to add 8.96 exaFLOPS to process new digital workflows, and are expected to increase data storage needs by 15,635 exabytes annually. FLOPS stands for floating point operations per second.

- Data gravity, as measured in gigabytes per second, is expected to more than double annually across the EMEA, Asia Pacific (APAC) and North America regions through 2024. Data gravity has a 139% CAGR from 2020 to 2024 globally; the number is highest in APAC at 153%, followed by EMEA at 133% and North America at 137%. Four of the top six metros expected to generate the fastest growth from 2020-2024 are in the APAC region: in descending order, Singapore (1st with 200% CAGR), Hong Kong (2nd at 177% CAGR), Sydney (4th with 159% CAGR), and Tokyo (6th). Beijing, Seoul and Shanghai were also studied.

Digital Realty noted that:

= Singapore is a critical business and data hub for global enterprises in the APAC region, given its pro-business policies and diverse connectivity options.

= Hong Kong is an international financial and trade hub and connectivity gateway for global enterprises between APAC and the rest of the world.

= Sydney serves as a global business hub with a significant global enterprise presence, in addition to serving as a rich connectivity gateway.

= Tokyo is one of the largest economies in the world, a business and financial hub for global enterprises and drives a significant volume of data creation.

Robert Davidson, Senior Director for network strategy, Digital Realty said that the research findings provide a data-driven indications of choke points for customers that Digital Realty can solve through tailoring its Platform Digital solution. "Are we acquiring dark fibre engineered to optimise for bandwidth or latency? It may be very different in the Sydney market versus the Tokyo market," he said.

"This gives us a guideline to be able to the focus on what's going to be the limitation for our customers in that given space."

According to Davidson, Asia has very different connectivity infrastructure compared to the US or Europe. While the US and Europe are mainly terrestrial (dry) transport with rare subsea (wet) links, Asian countries are predominantly connected with subsea links. "We're highly reliant on our subsea (links) and those subsea (links) have a higher level of latency," he said.

He also noted that repairing subsea cables can take a long time, recommending that data be held at two to three different data centres in an Asian country, with backups at other locations for ideal availability.

Digital Realty is the largest global provider of multitenant data-centre capacity by overall sq m. The company supports 4,000+ enterprises and service providers by delivering the full spectrum of data centre, colocation and interconnection solutions. PlatformDIGITAL, the company’s global data centre platform, helps customers to scale digital business and efficiently manage data gravity challenges. Digital Realty’s global data centre footprint includes over 280 facilities in 47 metros across 22 countries.

Explore:

Find out more about data gravity

Digital Realty is working on version 1.5 of the Index, which will have more more detail around industries and include data not in 1.0, McCrory said.

*The Data Gravity Index methodology is based on the analysis of thousands of attributes of Global 2000 enterprise companies’ presences in each metro, along with variables for each metro, including GDP, population, number of employees, technographics, IT spend, average bandwidth and latency, as well as flows of data. Digital Realty conducted research between August 2019 and August 2020 and drew upon more than a dozen third-party data sources, ranging from the World Economic Forum and the United Nations to global consulting and market research firms.

**Data localisation refers to the need for locating data in specific places for legal, security or performance reasons. Cyber-physical integration covers the integration of IT, operational technology (OT) and the Internet of Things (IoT) systems.