What Makes an Internet2 Application?

In this section of the seminar, I'm going to talk about what Internet2 applications are, what they will do, and who will help make it happen.

One of the interesting things about the Internet2 project is how we are combining an engineering effort with an applications focus. Generally, the Internet projects of the past have been very focused on the engineering and infrastructure challenges of building a network. With the Internet2 project, however, we're looking specifically at the applications that will drive the development of new research and educational opportunities. Our hope is that these opportunities will in turn draw people into using these new Internet services.

One of the questions I'm often asked is, "What is an Internet2 application?" And there really is no strict definition. That is, I describe Internet2 applications first as those applications that deliver qualitative and quantitative improvements over our existing research and teaching environments, and second, these Internet2 applications will not run on the current Internet. That is, due to some requirement of the application, it is simply not possible to run the application today. However, there is no strict metric for bits per second or latency or other technical requirements that differentiates Internet2 applications from other applications. With an Internet2 application it's essentially "you know it when you see it."

The goals we have on the applications side of the project include enhancing current applications with new features so that they can take advantage of the technical advancements made with the Internet2 infrastructure. What we also want to do is to create new applications that previously weren't possible. We really want to tap into the creativity of the people on our campuses -- faculty, researchers, students, administrators. People who work everyday in a higher education environment can come up with ideas that we haven't even thought of.

So I can't go and tell you about a specific "killer app" for Internet2. Our role is to simply enable what others -- what people like you -- conceive of and bring forth upon this new network.

So in the Internet2 applications development, we're working on a broad range of applications across all of the research and education activities in higher education. Some of the example applications we've been looking at include interactive research and instructional collaboration that allows participants from multiple distributed sites to work with each other as if they were in the same room.

Another application area we have been looking at is real-time, sensor-based modeling and simulation. For example, an application that takes advantage of this technology is computer weather forecasting, where you're gathering data out in the environment and then computing that data to anticipate storms. We also are exploring the use of data mining applications. These are large scale, multi-site computations applied to solving some sort of specific problem, such as health care or business data or research of all types.

There are also applications we're not yet capable of running at all -- and in fact it may be many years before we have them widely deployed. An example of this type of application is tele-immersion. Tele-immersion is where you have multiple participants inside a virtual-reality environment that renders a synthetic world. And we can keep on going! All and any part of these applications is an example of what we're working on for Internet2.

Now, to build these kinds of environments and to deploy these applications, we need certain parts of the infrastructure to be readily available. We don't want to have to build these enabling components into every application, so we'd like to have a base set of building blocks for all application developers. In the past, we've all agreed to use the TCP/IP protocol suite, but we haven't agreed on ways of federating or bringing together other infrastructure needs. For example, we need security that encompasses the many levels of authentication and authorization. We also need directory services that allow people to navigate through the network and assist in identifying a variety of resources -- whether it's people, servers, you name it. We need time service, for coordinating time across the network; accounting services, to identify the resources being used by individuals and to potentially bill for them; and shared file systems, so that users can access data from any source. We need web caching, to improve the storage capabilities of the network, and common application-to-application links, so that we can distribute applications based on component building blocks. And finally, we need services that support advanced features, such as streaming video, or multicasting technologies that improve the efficiency of service delivery.

One way to achieve some of our aims in applications development is to engage the people who have the experience in developing the types of applications we envision. I'm not talking about technical know-how -- I'm talking about hands-on experience in research or education. We need to actually go to campus faculty and talk to them -- they're not going to come to us. We have to go and locate them on campus or at their discipline-oriented meetings and inform them about the various features and services we're providing through Internet2. We need to establish a value proposition for them. That is, we need to get them to ask themselves, "How will Internet2 make a difference to me as a faculty member in the biology department, as a musician in the School of Music, as a practicing physician?" Once we've established for faculty and developers the value of Internet2 applications, we can help them understand how they can build applications to take advantage of Internet2.

There is also a variety of new issues that we're going to have to face. We need to build more portable, inter-operable, and scaleable applications so that we can more easily establish knowledge communities across higher education. We also have to talk about middleware technologies. These are the enabling technologies of the infrastructure, like multicasting, which provides the ability to deliver a lot of information to multiple users without sending it as individual streams. We also want to provide support for Quality of Service, so that applications and application users can establish classes of service -- for example, differential levels of bandwidth, data rate, or latency. Security is also important, whether we're protecting intellectual property that we've created ourselves or that we've licensed and are making available through digital libraries. We need to pull all these pieces together and make the knowledge available to the faculty and researchers on our campuses so that they can begin to understand the value of Internet2.

One of the things that we need to gather from these campus developers is their requirements for applications, and again, this goes back to the integration of engineering and applications development within the Internet2 project. The application requirements are those that we pass on to the network engineers so they can build a network that supports the applications we want to deploy. Applications developers and networking engineers are working together to establish the parameters of Internet2.

For example, we need to establish how many bytes per second of each application are passed across the network, so that we can then figure out capacity requirements. We also want to be able to design the network in such a way that it minimizes latency, a critical need for interactive applications. And we want to gather information about how many users there may be per application, to get an idea of scaling, and how many applications per campus, to determine the aggregate bandwidth requirements coming from the campuses onto the shared networks.

One of the things we have found is that very few application developers have an exact sense of what their application requirements are. We first need to help them understand their specific requirements, and then we need to help them measure those requirements by providing the appropriate tools.

The Internet2 working groups will help us in these efforts. There are working groups that are focusing on the specifics of the engineering side of the project, working groups that are focusing on the applications side, and joint working groups. For example, there are joint working groups focused on Quality of Service issues, multicasting, security and IP version 6. These joint groups are especially important because the requirements for these areas affect both network services and application components.

Beyond the joint working groups, we have the application-specific working groups that cover both discipline areas, such as arts and humanities and health care, as well as the areas involved in application infrastructure, such as distributed computation and storage.

Our working groups operate in a variety of ways. Some are simply information-sharing groups that provide resources through e-mail lists or webpages. Others may be more active, like the storage group, which will try to put together some pilots and tests to explore models or strategies for network storage on advanced high-speed networks. The requirements for the various working groups depend upon the needs and interests of each particular community.

One of the major tasks of the Internet2 project is to set up informational demonstrations. We want to show off advanced applications developed by the member universities. This helps establish the advantages and possibilities of these advanced networks so that people might be able to go back and create some additional applications themselves.

We also want to highlight the Internet2 partnerships with industry and government. It is only through our partnership with industry, for example, that we'll effect the technology transfer back to the commercial marketplace and make the services we're working on today the commodity services of the future.

The Internet2 applications that we've been working on so far have been in the areas of collaboration, access to remote resources, and visualization. In the area of collaboration, for example, we've shown how upper atmospheric researchers have worked across the globe. In the area of access to remote resources, we've demonstrated how scientists can operate remote microscopes and telescopes as if they were in their own laboratories. And in the area of visualization, we have tried to show how large scale data stores can be made accessible and useful to users.

Our current goal is to promote more arts and humanities applications, and more of the most demanding instructional applications. The arts and humanities have a variety of very interesting applications that we want to bring to the forefront in order to make people more aware of how these advanced networks can provide support for researchers and faculty across multiple disciplines. The most advanced instructional applications will incorporate streaming video and a great deal of interactivity. We are very excited about the potential -- and look forward to seeing the many new applications yet to be envisioned.

About CREN

© CREN, 1999

Contact Us