It appears that Uber is ready to super-size its compute infrastructure. The ride-sharing and logistics specialist has hired industry veteran Dean Nelson, the former head of data centers at eBay, as Head of Uber Compute.
Nelson’s hiring hints at the massive growth of Uber’s data infrastructure. Nelson is known for directing some of the technology industry’s largest computing, storage and e-commerce challenges at eBay, with a relentless focus on reducing the cost and environmental impact of its data centers.
It also suggests that Uber may soon consider building its own data centers. Nelson has a long track record of building high-efficiency hyperscale data centers, first at Sun Microsystems and later at eBay.
Uber generates and manages massive amounts of data to support its real-time dispatch system that matches drivers with riders using mobile phones. That requires real-time tracking of drivers in every city, accessing mapping and databases that span a huge global platform. The company’s operations will becvome more data-intensive as it expands into autonomous driving. Last week in Pittsburgh Uber rolled out a pilot program for its self-driving car service.
Hyperscale Meets Supercomputing
“Uber Compute is where Hyperscale meets Supercomputing,” said Nelson in announcing his new position at LinkedIn. “This is one of the most compelling big data problems I’ve ever seen. Imagine the amount of data being generated from these self driving cars. That data needs to be collected, ingested, analyzed and acted upon rapidly. This is Artificial Intelligence and Deep Learning on a grand scale.
“Today I’m poised at the starting line of this technology race,” he added. “I’m honored to join the team and innovate like crazy!”
In 2015 Uber leased wholesale data center space in three major markets, according to data center real estate reports. The company previously purchased a small data center from Microsoft, along with other assets that supported its mapping infrastructure.
Uber’s thirst for data center space illustrates the emerging demand from companies in two data-intensive sectors – connected cars and the peer-to-peer “sharing economy.” As these sectors mature, the leading players will be generating huge amounts of data, much of which must be managed in real-time systems spanning large geographies.
Massive Real-Time Data Platformg
At the heart of all these services is a real-time data platform that manages a dynamic market that now spans six continents. The company now operates in more than 340 cities in 61 countries, and has grown to more than 4,000 employees, including an engineering staff of 900.
Uber is privately-held, and hasn’t said much about its data center network. While it apparently uses the Amazon Web Services cloud for some things, it operates its own data centers and also stores some data on its drivers’ mobile devices.
Free Resource from Data Center Frontier White Paper Library
Distributed and Centralized Bypass Architectures Compared
When designing a power protection scheme for data centers, IT and facility managers must ask themselves whether a distributed or centralized backup strategy makes more sense. Unfortunately, there is no easy answer to that question. Download the new white paper from Vertiv that explores the principle of centralized versus distributed bypass and applies it equally to standalone monolithic and integrated-modular UPS architectures.Get this PDF emailed to you.
Drivers’ mobile phones send updates every 4 seconds as they move around. That’s why the design goal for Uber’s geospatial index is to handle a million writes per second, according to a recent presentation on scaling Uber’s market platform.
Uber provides an early example of the potential for connected cars to generate enormous volumes of data, which in turn will drive demand for data storage and data center space. Auto makers are expanding the data offerings on new models, going beyond on-board GPS navigation to provide safety and security apps (like “fatigue awareness” and stolen vehicle tracking, and even “infotainment” options.
A History of Building Big
At eBay, Nelson built new data centers in Phoenix and Utah that harnessed the latest innovations in hyperscale computing and energy management, including the deployment of modular data centers using warm-water cooling, and the installation of Bloom Energy fuel cells as the primary source of power. Nelson has also been an advocate of liquid cooling as the next phase in high-density data center design.
Nelson has plenty of credentials to run high-density compute in wholesale data center space, which he has done for eBay in its relationship with Switch, which hosts data centers at the SUPERNAP campus in Las Vegas. It remains to be seen whether Uber will continue with wholesale space, which can be deployed quickly, or whether Nelson will once again be overseeing data center construction projects.
Nelson’s focus on construction is reflected in his founding of Infrastructure Masons, a non-profit group of data center professionals created earlier this year to advance industry innovation and best practices. Nelson says he will continue to direct the group in his new post.
“I see the community we started six months ago as a critical component to staying connected to the people and the technology in this space,” Nelson said. “We are stronger together. My new challenge at Uber will help me contribute even more to the group.”
Data Center Frontier is edited by Rich Miller, the data center industry’s most experienced journalist. For more than 20 years, Rich has profiled the key role played by data centers in the Internet revolution.
Leave a Reply