Upon my return from sabbatical, I was very interested in initiating a new research project and I spoke with Badri (B. R. Badrinath Rutgers University Associate Professor, Computer Science Department; PhD University of Massachusetts) about this challenging technology and its consequences for new software design. He was enthusiastic about the new possibilities for research. The DATAMAN Project was launched in early 1992. In its five years of existence, this project produced more than 30 papers, several software prototypes, and a book ("Mobile Computing", Kluwer 1996). In addition, numerous grants totalling about $5 million were acquired from DARPA and the NSF.

Explorer: Since it's beginning, how has the DATAMAN Project evolved?
Badri: Through it's 5 years of existence, the DATAMAN Project has investigated a broad spectrum of issues including the new concept of wireless data broadcasting (air-disks), energy efficient communication protocols (how to maximize battery life in communication and data access), and wireless internet protocols to deal with the mixture of wired and wireless links (ITCP performs 2-4 times faster than existing TCP).

We have initiated a research group, the DataMan laboratory (located in CORE 335) and have organized several workshops and meetings at Rutgers including, MOBIDATA in November 1994.

The critical push to the project was given by major grants from DARPA and NSF which we obtained in 1995 and in 1997.

Tomasz and I have recently been concentrating on the issues of a wireless Internet. The concern here is how to provide on demand wireless access to the vast resources of the Internet regardless of physical location. Tomasz has also initiated, along with Julio Navas, the work on building a family of geographic protocols which integrate physical location (GPS) with the current IP based suite of protocols.

In addition, Anup Talukdar (Graduate Assistant) and myself are working on building a network that can support voice conversations between moving users; i.e., an Internet cellular telephone.

Explorer: What are the key challenges?
T.I.: Imagine a Personal Digital Assistant (PDA), a palmtop sized extension of a "Walkman" which will allow users to communicate freely at anytime in any location while having access to the vast information resources of the Internet. This device would interact with multiple wired and wireless networks, run on batteries for 8 or more hours of uninterrupted use, have the capability of operating from a power source inside of a car or a building, and be equipped with the Global Positioning System (GPS).

The PDA may also have several "network interfaces" that would be able to communicate using the future personal information network (wireless, cellular) with both narrow and broadband communication links. This technology would require the development of new and innovative software to make it practical. Consequently, the focus of the DATAMAN Project is not on hardware, but on providing software for such a PDA. Such software has to be extremely adaptive to respond to the rapidly changing resources (bandwidth, energy sources, tarriff structure etc). The entire client/server paradigm needs to be revisited, possibly adding an extra "proxy" layer - the agent residing in the fixed network who would act as an intermediary between the mobile client and the server. Finally, mobility introduces an entirely new family of location dependent applications which have to be supported by a new generation of software.

Explorer: What are the future implications of the DATAMAN Project?
Badri: DATAMAN has contributed to defining future Internet standards through the inclusion of two of its projects, GEOBONE and Wireless Proxies within the prestigious GLOMO (Global, Mobile) program as so called Integrated Technology Demonstrations. Six national experiments were selected by DARPA to be part of this program including the two experiments which are led by Rutgers University and include also UCLA, the University of California at Berkley, the University of California at Santa Cruz, and BBN as collaborators.

We have a strong chance to substantially influence the design of protocols and software which will run on every laptop and palmtop equipped with wireless connection and the GPS card. These are millions of machines and the potential technological impact can be dramatic. We are very excited about it.

Explorer: What is GEOBONE and GEOcasting?
T.I.: The GEOBONE project is similar in its ambitions and scope to its influential predecessor, the MBONE Project. Just as MBONE provided network support for delivery of multicast messages, the GEOBONE network will provide support for delivery of geographic messages. Multicasting is the ability to send the same message to several members of a specific multicasting group more efficiently than simply sending the identical message to each member individually. MBONE is a very successful experimental network capable of routing multicast messages and is typically used to broadcast video and voice for meetings and conferences, and even musical concerts. GEOBONE will be a network that will furnish similar functionality to route and deliver geographic messages that are direct to specific geographic areas defined as polygons of coordinates.

One may view GEOcasting as multicasting to the user groups on the basis of their current location. Geographic messages can be received by any terminal with a location providing device (GPS or other radio beacon sensors) that indicates that the terminal is located within the scope of such a message. Thus a user could literally walk or drive into a message. Emergency bulletins, community announcements, commercial advertising, and military communications are some of the practical applications. For example, you are driving on a highway and there is a traffic jam at Exit 9. Messages could be sent to the areas near Exits 7, 8, 10, and 11 informing drivers of the impending tie-up. Another example would be that your local Congressman/woman could send a local message to his/her constituents and save paper and postage. This capability should not be confused with the personal wireless communication which exists today. In geographic messaging, any authorized user could receive a message only if their current position were within the geographic address or intended polygon of the message. Messages would be limited to all customers within a one mile radius or to customers within a mall who have pagers.

Since 1995, commercial use of the GPS has been rapidly growing and soon every car and possibly every computer will be equipped with a location providing device. Broadly speaking, the goal of the GEOBONE project is to provide the entire software layer and the protocol suite to support GPS and other location providing devices. With DARPA support, we have a high chance to have a major impact in defining the future standards in this area.

Explorer: Have there been any new developments within the GEOBONE experiment?
T.I.: My doctoral student Julio Navas and I have designed the set of addressing and routing protocols to deliver geographic messages to locations specified as arbitrary polygons. This specification has recently been published as Internet's RFC 2009 with Julio implementing the first version.

The immediate plans of the DARPA experiment calls for the deployment of an experimental network with GEONODES at BBN, CECOM (US Army), University of California at Santa Cruz, and UCLA. The initial demonstration at the PI meeting in Los Angeles this February was extremely successful. In fact, another demonstration where an attempt will be made to deliver messages to targeted parts of Santa Cruz beach has been planned for July.

Julio and I are also initiating a new working group at Internet Engineering Task Force (IETF) which will hopefully lead to Internet standardization in the future. This means that the routing protocols developed at Rutgers could become a part of the standard TCP/IP based protocol suite of the Internet and hopefully a part of every machine which is equipped with a GPS card. If this takes place, a major dimension will be added to the logically addressable Internet. That dimension will be location. This will lead to an entirely new spectrum of WWW services such as distance based links and bookmarks where you could ask for all gas stations, restaurants, and shopping malls within a one mile radius of your location and distance based multicast where advertisements could be broadcast to all devices within a two mile radius.

Future goals of the GEOBONE project include providing the application programming interface to create applications which are location dependent and driven; and complete network support for geographic messages. The GEOBONE project has already attracted the attention of both the military and commercial sector. This exemplifies the notice that the DATAMAN project has drawn in its few years of existence and only in its second year of DARPA and NSF funding.

Explorer: What are Wireless Proxies?
Badri: The second experiment considers the integration of web proxies developed at the University of California at Berkley with wireless network adaptation fabricated as part of the DATAMAN project into adaptive software. The key challenge here is to adapt the network services to the rapidly varying conditions and changing resouces on the "last mile" of the Internet, as this link is often called.

The joint UC Berkley/Rutgers University demonstration at the PI meeting in February illustrated how expensive web pages could be adapted to narrow wireless links. In this way, the quality of the web access can be controlled through the concept of web proxies and adjusted to varying wireless conditions. This project is critical in building the future "wireless Internet" so that all Internet resources could be accessed from wireless platforms regardless of location.

Explorer: What other projects are you working on now?
T.I.: In addition to the two national experiments, Badri and I are collaborating with Richard Frenkiel and David Goodman from WINLAB on a new project called INFOSTATIONS. INFOSTATIONS are islands of high bit rate wireless connectivity: an information "cash machine". There would be "drive through", "walk through," Infostations depending on the mobility of the users. This new architecture leads to new communication behavior where a high quality channel is available for only a short period of time. The most recent piece of good news is that the Infostations project has been funded by DARPA, within its GLOMO-2 program with the new grant for over $2 million dollars for the 1997-2000 time frame.

Badri: We are also working on providing support for real-time services in a mobile internetwork. Internet cellular phone being one of them. We are also working on providing support for disconnected operation for databases. Support for disconnection allows users to hoard portions of data from a database server, and work on the data while being disconnected.

For more information on the DATAMAN project, see:
http://www.cs.rutgers.edu/~badri/dataman

Open House

Conversing

Data Mining

Mobile Wireless Computing

Women in CS

Remote Learning

PhD Recipients

Recent Awards

DCS Web Page