Thursday, June 13, 2024

PlanetLab ‘Slices’ Into Alternate Internet Universe

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

A consortium of universities and high-tech companies Tuesday formally launched a joint Internet-based test-bed platform for developing massive online services using distributed computing and overlay networks.

The project, dubbed “PlanetLab” is hosted by Princeton University but is a grass roots effort currently made up of 60 different academic institutions with 160 machines at 65 sites around the globe. The goal, say organizers is to make it so that educational organizations and private enterprise can build applications for the next generation of the Internet without bringing down the current system we know and love today.

”[The Web is] so successful and so many people depend on it, it’s become impossible to go to the core of the Internet and make radical changes to introduce the kind of new services we see people wanting to deploy,” Princeton University scientist and Intel Research member Larry Peterson said during a conference call to the press.

The short-term goal is to attach upwards of 1,000 widely-distributed machines in the next two to three years.

To date, more than 70 research projects at top academic institutions including MIT, Stanford, UC Berkeley, Princeton and the University of Washington have used PlanetLab to experiment with such diverse topics as distributed storage, network mapping, peer-to-peer systems, distributed hash tables, and distributed query processing.

Peterson says high-tech companies will now be key to the project’s success. Already Intel has provided seed money and donated 100 computers. Hewlett-Packard followed suit with 30 more computers for 10 additional sites. More tech companies are expected to come on board in the next few months.

“This is about pooling resources and to build out the infrastructure, but in the end this about lowering the barrier to entry to developing on the Internet,” Peterson said. “One could imagine that there could be many public and private Planet Labs that co-exist and interoperate between each other.”

Different than the Internet 2 project or even Grid computing, the group says the most obvious benefit is that network services installed on PlanetLab experience all of the behaviors of the real Internet where the only thing predictable is unpredictability (latency, bandwidth, paths taken). A second advantage is that PlanetLab provides a diverse perspective on the Internet in terms of connection properties, network presence, and geographical location. The broad perspective on the Internet enables development and deployment of a new class of services that see the network from many different angles.

The PlanetLab software is based on the most current release of Red Hat . Organizers say the Linux operating system was chosen because of its ability to get up and running fast and handle extremely large amounts of computing.

Another key objective of the software is to support distributed virtualization — the ability to allocate a portion or “slice” of PlanetLab’s network-wide hardware resources to an application. This allows an application to run across all (or some) of the machines distributed over the globe, where at any given time, multiple applications may be running in different slices of PlanetLab.

“This is just the beginning of a new class of services and applications that are distributed over much of the Web and will affect the design of intelligent servers, network storage and network processors,” said U.C. Berkeley computer science professor and co-director of Intel Research Berkeley David Culler.

Two current examples include UC Berkeley’s OceanStore, which is global persistent data store distributed over the whole Internet and Intel’s Netbait, a research platform that detects and tracks Internet worms globally.

Other projects onboard include the CoDeeN content distribution network (Princeton); Sophia distributed query processing engine (Princeton); PIER distributed query processing engine (Berkeley); ScriptRoute network measurement tool (Washington); Chord scalable object location service (MIT, Berkeley); and SplitStream multicast streaming system (Rice).

“What we are doing is a circle of design, deploy, and measure,” said Culler.

The next step, according to Culler, is for participants to set up a consortium that will operate the network and to invite other companies and education institutions to participate. The majority of the sits are located in the United States, Europe and Australia. The group says Russia, China and India are being approached to add to the mix.

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles