Report on Future Trends in Networking Technology Networking and Telecommunications Standing Committee Council on Information Technology and Services University Of Florida December 6, 1994 Prepared by R. Smith and R. E. Newman-Wolfe CSE-346 392- 1488 nemo@cis.ufl.edu http://www.cis.ufl.edu/ nemo Report on Future Trends in Networking Technology Networking and Telecommunications Standing Committee December 6, 1994 1 Status of This Report As part of the Network '99 planning process NTSC was asked to provide a report on Future Trends. This appears as · Recommendation 1. NTSC should create a working group to appraise emerging technologies and their likely impact and value for UF networking. This report should evaluate the desirability and feasibility of each technology. The report would address specific networking goals or statements of direction from the Council for the following year and would build in an annual review process and updating of the document. This report is due in the office of the Executive Director by July 1. This report was approved by the Networking and Telecommunication Standing Committee (NTSC) of the Council on Information Technologies and Services (CITS) on November 22, 1994, pending the changes present in this version. It will be presented to CITS on December 8, 1994. This report (and others from NTSC) are available via gopher at gopher://gopher.cis.ufl.edu/admin/ntsc/documents/network-trends.ps. The relevant Network '99 recommendations are quoted below. 2 Future of This Report As part of the annual review process, NTSC should review the " Future Trends" document from the previous year early in the spring. This would allow Network Services and other campus organizations involved in networking to incorporate any recommendations into their budget request for the coming year. However, no emerging network technologies will be a factor at UF until a consistent and adequate funding base can be established for maintenance and growth of this mission-critical resource. 3 Introduction Networking technology is changing rapidly in almost all areas. Decisions about which of the emerging technologies will ultimately prove useful in the UF environment, let alone hold enough market share to be well supported and economically feasible cannot be made at this time. Historically, UF has made technological advances in information technologies either through close vendor relationships or by waiting until commodity level pricing was reached. Many of the problems of being at the cutting edge of technology are eliminated as a result. In this report we will evaluate the various technologies on the horizon with an eye for how they might impact UF. There are three general directions that we are watching: mobile computing, client-server computing, and multi-media applications. Some applications, like Mosaic, combine two or more of these modes. Client-server and multi-media both tax the available bandwidth of the wired network. Some forms of multimedia like interactive video require low latency to be effective. There are various ways available now and coming available to handle the bandwidth problems. Whenever we are dealing with new technology, we should follow standards which come from an open discussion whether that be the Internet RFC process, the IEEE, the ISO, or NIST. 4 Technologies In what follows, we outline the advantages and disadvantages of various technologies and approaches that are likely eventually to play a part in networking at UF. 1. Speeded-up existing technology. Ethernet, Token Ring, and FDDI all have speeded-up versions which are in the standards process and in some cases available on the market. · Advantages: (a) Directly addresses the available bandwidth problem. (b) With luck, it is not necessary to change the cable plant to utilize the speeded-up versions. (c) Good software drivers should be available. (d) Network management should be essentially the same for the speeded-up technology as its slower predecessor. (e) Relatively inexpensive. · Disadvantages: (a) Other than bandwidth, the disadvantages of the old technology are still present. (b) One does not imagine that these solutions will scale up too much more, so that good planning views these as temporary fixes. We expect these approaches to be scale for anticipated growth for the next five years, except possibly in the core network and special subnets. 2. Switching hubs. Ethernet, Token Ring, and FDDI all have switching hubs available. · Advantages: (a) Almost a plug 'n play technology, pull out the old hub, plunk in the new switching hub. (b) Relatively inexpensive (price coming down, less expensive than ATM). (c) Same software drivers should work. (d) Some of the switches offer security filters. · Disadvantages: (a) The performance gains are tied somewhat closely to the traffic patterns of the network. Further tweaking of the topology may be necessary to realize a gain in performance. (b) This solution does not scale up, so that good planning views these as temporary fixes. Switching hubs are best suited for LANs where there is both the cable plant and the traffic patterns needed. 3. Integrated Services Digital Network. Regional operating companies, including Southern Bell, have begun to offer ISDN. · Advantages: (a) Tarriffed and widely available (b) Higher speeds than modems · Disadvantages: (a) Not all that fast (see competition from cable companies) (b) Not cheap locally (c) Somewhat limited coverage We see ISDN providing for access to the home, for instant infrastructure (particularly for remote locations in town), and for video-conferencing applications. 4. Asynchronous Transfer Mode. Since this is a cell based switching method, it might rightfully belong in the discussion on switching hubs. Unlike the switches discussed above, it is not based on existing protocols. The key features of ATM are: · Advantages: (a) A small cell size which permits "multiplexing" of signals from many sources. (b) The design objective was to permit voice, video, and data to all flow on the same network - a multi-media technology. (c) Bandwidth allocation to a user and an application. (d) A fast, hardware switch based network. (e) Strong vendor support. No sane vendor is without an ATM strategy these days. (f) The ATM protocol is designed to scale with the SONET protocol to very fast speeds (100's of Mbps). (g) The intent is that this will be the single WAN, MAN, LAN protocol. This could greatly simplify wide-area, desktop to desktop communication. (h) Better security. · Disadvantages: (a) So far, only one protocol, TCP/IP, is being supported (though Novell IPX has recently been claimed, and most protocols of interest are likely to be supported by the time we can afford it). (b) Rather few vendors are talking about routing in the ATM network. There are no routing standards established yet. (c) So far, only 155 Mbps speeds are supported. Until ATM is proven to go faster than this, the advantages over FDDI are not so apparent. (d) The telecommunications carriers will use this technology to reinstate measured rate service for the Internet. (e) The hype is oppressive and way beyond what is deliverable. (f) Still working on broadcasting to support LAN protocols. ATM is likely to be used first by units with particular needs and budgets internally, much as FDDI is currently used by similar units. Eventually, the UFNet backbone may migrate to this technology. 5. Cabling Media. - Unshielded Twisted Pair. UTP cabling is proving to be remarkably resilient. · Advantages: (a) Cheap (b) Easy to install. (c) Structured wiring system. · Disadvantages: (a) Acts as an antenna for transmitting and receiving RF noise. (b) Limited bandwidth. This is likely to be fine for use at the unit level for the next five years. It is basically economical and fairly easy to install. Some of the newer versions appear to be able to carry 300 Mbps or more. Much depends on timing and vendor interest in supporting UTP at the higher speeds. This is not a media for electromagnetically sensitive environments. UTP is a horizontal cabling. - Multi-mode Fiber. · Advantages: (a) Cheap (as fiber goes), easy to install (we have the tools, experience). (b) Large installed plant (c) More secure than metal · Disadvantages: (a) Limited distance. (b) Limited bandwidth (compared to single mode). (c) Opto-electronics still expensive. Fiber (of both types) is the choice for high-speed LANs, the campus backbone, for video applications, and for harsh environments. - Single-mode Fiber. · Advantages: (a) Tremendous distances. (b) Tremendous bandwidth. (c) More secure than metal · Disadvantages: (a) Yet another termination procedure. (b) Almost no installed plant. (c) Opto-electronics still expensive - Coax (the amazing cable that refuses to die). The cable television industry is staging a revival of data-on-RF cable. · Advantages: (a) Amazing bandwidth. (b) It's everywhere you want it to be. (c) A multi-media solution. · Disadvantages: (a) Spaghetti network topology. (b) Security? (c) Two-way cable RF systems are a pain to maintain, fortunately we will not have to, but we will probably be blamed for the problems. (d) Network management? (e) Routing? Coax is still useful for broadband video, and for access to outside network providers (e.g., for access to the home through cable companies). It is not recommended for on-campus data transport. 6. Wireless. Without a doubt the future of networking will have a serious wireless component. With the FCC auction of the Personal Communication Services (PCS) bands this year, as well as the various mergers in the wireless communications industry, we will see a large variety of wireless product emerging. There are already many wireless products in the LAN market. · Advantages: (a) Mobility (b) Highly personal, user supported device. (c) Offers the possibility of a device that students will have when they arrive, use while they are here, and take with them when they are gone. · Disadvantages: (a) No standards. A bewildering variety of LAN protocols available now. (b) Highly personal, user supported device. (c) Roaming protocols? (d) Network management? (e) Low bandwidth. (f) Health risks? Uses of wireless include fast network restoration after outage, and mobile computing. 5 Conclusions It is very difficult to guess how many people will drive across a bridge by observing how many swim the river. In order to make any reasonable projections of needs and appropriate technologies, the user population must be surveyed. This includes the faculty, administration, staff, and students, who should be asked about both desired uses (teaching, research, service, management, work at home, etc.) and requirements (speed, distance, reliability, delay, costs, etc.). Until this survey is made, it will be difficult to determine the impact or the value of deployment of emerging and often expensive technologies.