« Previous Topics June 2011Future Topics »

National Research and Education Networks

The State of Research and Education Networking in Africa

African NRENs can expand educational opportunities across education sectors

Louis Fox

Over the past three decades, the revolution in computers and telecommunications networks has created unprecedented changes in business, commerce, government, science, health care, and education. New jobs, new industries, an explosion in entrepreneurship, new modes of community building, increased learning opportunities, ease of access to timely information and global markets, and the ability of an extended community to interact closely across space and time: all are dividends of this revolution in network and information technology and the remarkable underlying Internet culture of change.

Yet the fruits of this Information Age are still unevenly distributed. This gap threatens to continue to cut off some populations from new opportunities. Access to new forms of education, good jobs, medical and health information, communication, and the chance to participate in the affairs of the broader society may be denied to them. For some individuals, technology brings the promise of inclusion, education, opportunity, wealth, and better health; for others, greater isolation and continuing poverty. Many look to universities and K-12 schools to bridge this gap.

Meanwhile, today’s Internet—the commodity or commercial Internet—has recognized a number of limitations. At the same time numerous opportunities and new possibilities have emerged. Some challenges, like the inability to provide workable “quality of service” or end-to-end performance guarantees needed for demanding applications such as telepresence (the current state-of-the-art videoconferencing technology) were outside the scope of the Internet’s original design goals. Challenges, such as dealing with today’s gargantuan amounts of traffic, exploding number of users and sites, privacy and security needs of users and institutions, and requirements for Internet addresses, are the consequences of unanticipated success.

Many new but challenging opportunities, like the delivery on demand of real-time, movie-quality, high definition television (HDTV) or even films over the Internet, as well as many new and experimental approaches to health care, are the product of extraordinary progress in a wide array of technology industries that are now convergent with the Internet’s evolutionary path. Other new applications and capacities are outside the focus of the commercial Internet. These innovative activities are supported best by research test-beds, the international fabric of national research and education networks (NRENs), which focus on the development and deployment of the next generation of Internet technologies.

The regular or “commodity” Internet was not designed to handle the huge amount of data transfer, the explosive numbers of users, or the interactive, media-rich applications commonly used today. For applications where reliability is critical and delay is unacceptable – applications such as real-time streaming events, access to remote scientific instruments, high definition video-conferencing, online gaming, and interactive immersive worlds and simulations – the commodity Internet is inadequate. Research and education networks were purpose-built by the research and education community to offer the flexibility, performance, speed, and advanced services that allow these applications to evolve and thrive.

NRENs serve many functions. They create leading-edge network capability for the international research community; they enable revolutionary Internet applications; they ensure the rapid transfer of new network services and applications to the broader Internet community; they provide a platform for sharing scientific (and other) applications and resources; they aggregate demand for bandwidth and thereby create “buying clubs,” drive down the cost of bandwidth; and they create social value by including communities outside their primary research university constituencies, like primary and secondary schools, libraries, museums, scientific and cultural institutions. In order to flourish, NRENs must focus on the technical dimensions of data networks and they must also attend to the human dimension, the creation of shareable expertise for support and collaboration across many fields of research and education.

The African Context for NRENs

NRENs began in Africa about ten years ago, with Eastern and Southern Africa at the forefront. The availability of fiber and the high cost of bandwidth were, initially, limiting factors. Now, with several trans-oceanic submarine cable systems completed or near completion, and with a concurrent expansion of terrestrial fiber across Africa, access to fiber is within reach on most of the continent. Prices have dropped significantly, although bandwidth is still pricey when compared with rates in many other parts of the world. NRENs can help to address pricing inequities across countries by (a) aggregating demand among universities and, more broadly, within the school sector (more on this below); (b) architecting networks with points of presence across broad and complex geographies; and (c) and by working across national boundaries to create regional optical networks and, ultimately, a pan-African optical network.

Furthermore, African NRENs can leapfrog their counterpart NRENs elsewhere in the world and build networks without some of the inherent historical limitations of comparable networks, emphasizing collaboration and mass access to education and research applications across educational sectors. In addition, African NRENs can design their networks to combine the best of wireless and mobile technologies with optical networks. Inspiring leaders, ambitious goals, and imaginative and carefully crafted plans – these things (and more) will guarantee that African NRENs will flourish.

The continent has a firm foundation in place. There are NREN success stories such as KENET in Kenya, RENU in Uganda, TENET in South Africa, Xnet in Namibia, to name a few. And there are regional efforts, the most prominent of which is the UbuntuNet Alliance, which began as a regional bandwidth aggregator and now has created a very strong human network and an operational point-of-presence which can, over time, be the initial hub of a regional network. The UbuntuNet Alliance is, in fact, a model for subsequent developments in West Africa (WACREN) and North Africa and the Arab States (ASREN) – both of which are nascent regional networks, now human networks and, eventually optical networks.

The cornerstone of the R&E networking is the Local Area Network (LAN), which is the network serving a university, school, museum, or research institution, and the network closest to the end-user. In some instances, these LANs might connect to a municipal network or another Wide Area Network (WAN) and then to an NREN. In other instances, the LAN may connect directly to the NREN. Similarly, NRENs may connect to a multi-national regional network or directly to other international NRENs or, perhaps, to a pan-African R&E Network. Much will depend upon local conditions, regulatory structures, and geography. (In its ideal state, networking is a function of the best technological practices and geography, not politics.) Figure #1 below illustrates the various strata of networking.

NRENs: A Necessary Foundation for African e-Science

Advanced information, communication, computation and collaboration technologies – known as cyberinfrastructure – have become essential elements for education and for research in the 21st century. Of particular interest to many researchers and educators is the use of these tools for “e-science,” as computational discovery has emerged to complement the traditional practices of theory and experimentation. Examples abound across all scientific disciplines, as well as in the arts and humanities.

Explosive growth in the resolution of sensors and scientific instruments has led to unprecedented volumes of environmental and experimental data, which can be combined, compared, and correlated across time, place, and types of data. Computational science aids in modeling, simulation, and scenario assessment using data from diverse sources. Complex multidisciplinary problems – from health care and public policy to national security, scientific discovery, and economic competitiveness –complement the historical focus on single disciplines. And important multidisciplinary discoveries are now made by teams of experts spread around the world.

Advanced cyberinfrastructure, enabled by very high-speed research and education networks, is essential for participating in all these efforts. Those without access and the ability to participate will not have full participation in 21st century innovation.

Therefore, a major challenge confronting African nations today is how to ensure that all colleges and universities, including those that have not traditionally benefited from expensive research infrastructure, can participate seamlessly in national and multinational e-science efforts that are cyberinfrastructure-enabled. The challenge begins with the need for ubiquitous deployment of advanced research and education networks.

NREN Practices to Consider

Peering
As the Internet evolved from a US government funded network in the 1980s to a world-wide, market driven network in the 1990s and beyond, one organizing principle continues to endure – the settlement-free exchange of Internet traffic among independent networks. Often referred to as “peering” by the community of engineers and operators of networks, this seemingly contradictory notion of the free exchange of traffic among competitors as an economic benefit has become an important foundation in the growth of the network. Large centers of settlement-free peering have also resulted in greater network resiliency in light of geographic or systemic outages, and the promotion of fair and equitable access to the constantly evolving Internet marketplace.

There are a few key structural principles one may wish to consider when implementing settlement-free peering facilities in an emerging NREN or regional network:

• Geographic diversity. Internet routing decisions often follow the “first exit rule”. Thus, a network needing to pass data to another “peer” network will usually pass that traffic to its peer at the first opportunity. This often results in networks only agreeing to peer with one another if the peering facilities are distributed widely in a given geographic area.
• Resiliency. One should build a high degree of redundancy in all of the necessary components comprising a peering facility; electricity (i.e., multiple feeds with generator backup), diverse fiber paths in to and out of the facility, and “carrier class” environmentals such as HVAC, security, and fire-suppression.
• Open access. A peering facility should have equitable, open and easily understood criteria for all participants who wish to connect to the peering fabric (i.e., switches, routers, fiber-distribution panels). The more participation from networks in a peering facility, the higher the degree of usefulness to all concerned.
• Sustainability. The success of a peering facility itself becomes a potential service liability if the facility is underfunded or inadequately maintained. Early peering facilities in the US in the 1980s were sponsored and subsidized by the federal government, with commercially managed peering facilities quickly following once economies of scale were reached. Depending upon the financial realities of a nascent deployment of continental peering facilities, one may want to consider government subsidy and oversight of early peering facilities until an economy of scale is achieved to allow a more independent yet still reliable support model.

IPv6
Conventional computers have been joined on the Internet by a myriad of new devices, including iPads and smart phones, smart TV set-top boxes and videogames with integrated Web browsers, and embedded network components in equipment ranging from office copy machines to kitchen appliances to automobiles.

Internet Protocol version 6 is needed because the Web is running out of addresses. The current technology, known as Internet Protocol version 4 (IPv4), supports just 4 billion addresses, not nearly enough to cope with the new devices that connect to the Internet and need addresses and certainly not enough addresses to cope with the explosion of new devices across the African continent.

With the future in mind, IPv6 has been outfitted with an enormous address space that should provide globally unique addresses for every conceivable variety of network devices for the foreseeable future (i.e., decades).

But IPv6 is a complex structure and addressing is only the most visible component. IPv6 also attempts to deal with critical business requirements for more scalable network architectures, improved security and data integrity, auto configuration, mobile computing, data multicasting, and more efficient network route aggregation at the global backbone level.

Middleware: Access and Identity Management
The term “middleware” is used to cover a broad array of tools, information, and what programmers call “hooks” that help applications use advanced network resources and services. Middleware can be thought of as glue layers that provide reliable, standardized support services like authenticating users and authorizing them (or not) to use specific applications or have access to certain on-line resources. Indeed one common application of middleware is to provide the common services and information necessary to allow applications to restrict or enable access (“log on”) to certain resources.

Middleware such as authentication (are people or programs who they say they are?), authorization (what is he/she/it allowed to do?), and the directory services needed to keep track of users, resources, and any rules that may apply to them, comprise essential elements of any shared network computing infrastructure. Other middleware services, such as cooperative scheduling of networked resources, enabling secure multicast or interactive video or object brokering (matching requests with providers for relatively high level services, such as databases, format, or protocol conversion) are preconditions for many applications and services sought by the research and education communities. These include a number of innovative applications.

Broad adoption across education of certain standardized middleware fabric is a key requirement for addressing the needs of the education community for capabilities like user-friendly, but broadly shared and highly cost-effective access to libraries and other educational resource repositories, remote scientific tools, music repositories, and other intellectual property; for use of widely and safely shared interactive services; and for workable and properly protected wide-scale student records access and transmission. As such, middleware must be, as a practical matter, interoperable between applications, among campuses and other educational institutions, and the wider Internet. This effort will not be successful if individual groups or institutions build their own internal versions of middleware and then try to patch the pieces together. African NRENs are at a distinct advantage here as the compromises required to develop a common framework, standards, and protocols for attribute naming, storage, and exchange are easier to obtain when there are no existing use cases.

However, developing and managing the trust relationships necessary for the success of identity management can be tricky. The more diverse the groups, the more complex this becomes, particularly when the focus is inclusion of many educational sectors beyond universities. One should expect significant challenges as divergent interests and priorities will be even greater in this environment. The bottom line is that the technical issues are the least difficult to address. New policies specific to access identity management, and the operational issues caused by them, tend to be bigger hurdles. As with introduction of any new processes, effective change management will play a significant role in successful outcomes.

Some engagement of organizations like UbuntuNet and key leaders among existing African NRENs in international access and identity management federations like REFEDS would, ultimately, be extremely beneficial to successful implementation of middleware across diverse educational sectors among these NRENs.

Wireless Access
Given the prevalence of mobile and wireless technologies for mass access to education in African countries, careful attention to the integration of the various forms of wireless technologies – Microwave, Wi-Fi, WiMax, and cellular (3G, 3.5G and 4G) – is critical. These are all excellent ways to extend the reach of wired R&E networks. The best practices are dependent upon the environment, potential commercial partners, available spectrum, and other local conditions.

Wi-Fi is still the leader in terms of network speed. It is best suited for building or campus environments. The equipment is inexpensive and readily available. WiMax and cellular networks are usually deployed in connection with a wireless service provider, although there are several examples of communities and institutions deploying their own. The real differences between 3G/4G are data-rates and the amount of spectrum that is in use. For instance, 3G networks can exceed the speed of a T-1 line (a fiber optic line with a 1.5Mb/s speed). Second generation data networks (2G cellular) still have a place as they are widely deployed and their slower speeds often mean less cost.

Extending the Reach of African NRENs: Supporting Schools and SchoolNets

NRENs can provide significant social benefit by extending their reach to schools and other educational institutions (e.g., libraries, museums, scientific and cultural organizations). Such efforts can contribute to the development of prospective university students who can begin to develop fluency with information technologies while in primary and secondary schools. In addition, there are many compelling models of university students being trained to be both technology and content experts who intern at school sites and in doing so, enrich their own experiences as well as the students and teachers whom they support. It is a wonderful way to train students, particularly those in non-technical fields who may aspire to occupations where information technology is either at the center of their work or essential to it.

In the U.S., the K20 Initiative now engages schools in 43 of the 50 states, and over 70,000 schools and millions of students. It was not conceived at the outset of the creation of Internet2 but has become one of the hallmarks of the U.S.’s advanced R&E network initiatives. If African NRENs are essentially greenfield efforts, extending their reach to schools would have many benefits. By increasing the numbers of institutions participating, such an effort could have a positive impact financially by aggregating bandwidth costs across significantly more institutions.

Broadly stated, a schools initiative can have many goals, which may include the following: (1) to bring innovators in K-12, colleges, universities, libraries, and museums into appropriate regional, national, and international advanced networking efforts, creating new “workgroups” where warranted; (2) to develop mechanisms for enabling quick, pervasive technology diffusion and transfer; (3) to create mechanisms for timely communication across educational sectors and regions; (4) to leverage and propagate a culture of parallel independent efforts along with education, private sector, and government partnerships; (5) to get interested and capable schoolnets connected and properly engaged in existing workgroups and projects; and (6) where there is interest and realistic opportunity, to include appropriate experiments in learning and education and help enable experiments involving innovative deployments of advanced technologies in education at school sites.

Among the many activities of such an initiative, relevant local, provincial, and national special interest groups might be formed in some of the areas described below to pursue collaborative ventures:

• Digital learning resources, content repositories and open educational resources
• Learning management systems and education management systems
• Videoconferencing: H.323 and other interactive video and multimedia technologies, digital video, low- to high-end video multicast, and the convergence of on demand video and broadcast
• Access to scientific apparatus and other broad application areas which could be shared across educational communities
• Middleware, enhanced portal, and “relationship-ware” deployment and partnerships
• Advanced server technologies, caching, and co-location strategies
• IPv6 deployment
• “Buying clubs” to purchase access devices (computers, mobile devices, etc.)
• Cloud resources

Despite the many challenges and complexities ahead, African NRENs have innumerable opportunities to expand educational opportunities across the widest range of education sectors, to create a platform for African faculty and students to engage in research collaborations across the continent and the globe, and to support a rising generation of researchers, educators, professionals, and leaders who will contribute to a peaceful and prosperous Africa.

This discussion is part of the eTranform Africa initiative.

Written by Louis Fox on June 27, 2011 in National Research and Education Networks.
Tagged: African NREN ASREN cyberinfrastructure HDTV Information Age IPv6 KENET Middleware NREN Peering RENU telepresence TENET UbuntuNet Alliance Xnet | Trackbacks