First published in AMS business magazine. Author: Paul Anstiss

As traditional databases have become swamped by all the data being produced, more companies than ever before are joining the inexorable migration to the Cloud for their storage needs. However, this comes with a cost. Lago points out that, not only does the data sector consume a great deal of energy to keep systems cool and regulate humidity, it has a large capacity that is currently not being utilised efficiently. Currently, it’s estimated that the sector consumes 10% of the world’s electricity, and within 20 years it is expected to become the main consumer. 

Lago’s research is concerned with software- and service-oriented architecture, architectural knowledge management, and green IT. She and her team see themselves very much as pioneers in an industry that is only just beginning to wake up to the possibilities of big data. At the VU, Lago and her team are building new energy- and sustainability-aware software architecture that allows data storage systems to ‘sip’ energy instead of guzzle it. “The impact of this migration to the Cloud is huge, and the business world doesn’t always appreciate what’s involved,” says Lago. “It tends to focus on privacy and security issues because we understand these better, but we also need to consider the availability and reliability of these systems, not only for safety-critical services in case of disaster, but also for the services we use in our daily lives, such as banking.”

Amsterdam hub

It is no coincidence that Lago and her team find themselves at the forefront of this brave new world. Amsterdam has become the digital gateway for all Internet traffic arriving in Europe from the US, and one third of all of Europe’s data centres are located here. The region’s cool climate and flat landscape make it the ideal environment, and in the Amsterdam Randstad area, there are
180 ‘data hotels’ covering more than 240,000 m2. Lago describes it as a cascading effect that’s seen companies such as IBM, EMC and Google look to the Netherlands for their data-storage needs.

“Maturity in information and communications technology design, together with a government strategy that actively invests in key sectors, has put the Netherlands ahead of other countries in this industry, despite it being so small.”

Green lab

Lago says companies often decide to outsource their data management for economic reasons without actually thinking about the long-term costs, but she says it’s a necessary consideration when designing the architecture of a new system. Not only should a data storage system be green, it also needs to be profitable and designed for change. According to Lago, the first thing to establish is what data should remain in the company and what should migrate to the Cloud.

At VU’s Green Lab, Lago and her team are helping design a Cloud model optimised for company needs. Collaboration with the world of industry is seen as the key to success. The lab hosts a number of servers with state-of-the-art energy sensors, to run experiments on software energy efficiency. The aim of the research is to focus on the real needs of industrial practice. Researchers look for the ‘software footprint’ that makes software smart as well as sustainable; they measure, for example, energy consumption and try to find new ways to optimise behaviour. “We focus on the energy efficiency of the systems themselves, which means we can lessen the amount of data stored and decrease duplications” says Lago. “We need to decrease the number of servers managing data that no one uses and look more broadly at how software systems managing processes like transportation, logistics and mobility can be organised in a more sustainable way. This has a direct impact on the amount of servers involved and the amount of electricity that they use.”

Smarter systems, greater optimisation

The VU is world renowned for its creativity in software engineering. In particular, it focuses on architectural tactics that make software systems smarter and provide a better quality of service. Furthermore, the team of researchers is looking at ways of turning the disadvantages of a system into advantages. For example, in several quarters in Amsterdam, the heat generated from data storage is used to heat residential areas. In fact, Amsterdam’s City Hall was among the first to use the heat created from its data storage unit to grow tomatoes in its own greenhouse next door.

Although cooling is less of a problem than it was, the size of storage centres continues to grow. For Lago, the most important challenge is optimisation. “In recent decades, we have been developing software without any attention to optimisation, simply because hardware has become so cheap and efficient that no one worries. Systems have grown in size and store a lot of data that no one uses. But I believe that quality of service and flexibility will become a key differentiating factor in the future.”

The key to success

Lago is confident that her work at the VU remains at the forefront of developments in big data, but she does admit that it’s been hard to convince the busines world to provide the necessary funding for research. As Lago sees it, innovation is the key to the success of the big-data revolution, and the ability to accurately predict how much data-storage capacity is needed will determine the future direction of this exciting new industry. But she warns that many data centres, which currently use only 20-80% of their storage capacity, are not in a position to make the predictions they need to put things into context. According to Lago, this overcapacity is expensive and the potential for decreasing the safety buffer is immense. For big data to succeed in the future, the problem must be tackled now.

“Unfortunately, there is currently no incentive for companies to invest in better predictive technology, but in a few years they will have to. Those companies that invest now will find that they are ahead when it comes to providing data-storage and software services that require less energy and offer greater flexibility.” Lago’s predictions are a wake-up call to the data-storage and software industry: don’t let big data get bigger than your business.