The Connected Schools of Belen, New Mexico: A Wireless Success Story
This report is part of theinformation-gathering for the National Science Foundation's WirelessField Tests for Education Project. We are installing and testing,under a wide variety of rural and urban educational circumstances,various wireless devices -- from short-range wireless LANs, to 25mile T1 radios -- in Colorado. This report infull, and others pertaining to the project, can beaccessed at the Project's Web site: http://wireless.oldcolo.com.
The 8 schools of the Belen Consolidated School District of ValenciaCounty, New Mexico, span an area over 5 miles wide and 10 miles longcentered on this rural agricultural town of 6,500, some 30 milessouth of Albuquerque. All 8 schools are linked to each otherwirelessly, in ways that dramatically demonstrate the potential, andcost-effectiveness, for typical K-12 school network operations usingdigital radios operating in frequency bands approved by the FCC forunlicensed use. They also illuminate the limitations of using radiosthat are required to operate under Part 15 FCC rules, and point torequirements that cannot yet be satisfactorily met in the promisingfield of spread- spectrum digital radio. In essence, Belen schooldistrict is totally networked by no-license, no-monthly cost, T1wireless communications -- and linked to the Internet from allschools through this network.
Belen is rather typical of a rural district --serving an area with approximately 12,000 residents, half residing inBelen proper and the remainder on small farms in the surroundingarea. It has 6 elementary school buildings, 1 middle school, 1 highschool, and an administrative center co-located with an elementaryschool. It has 4,800 students, with a faculty of 250. The district iscomposed of lower middle- income residents; 72% of students areHispanic.
The town and surrounding area are situated on theflat shallow basin west of the Rio Grande River. There are no highhills, but the high school is situated on a rising bluff at thewestern edge of the district, giving a rough radio line of sight tothe other 7 schools. There are no significant natural or man-madeobstacles in town.
Districts in the U.S. have fourbasic needs
for no-license (or no-recurring-cost) wireless.
The school district made a major decision in 1994,under then-Superintendent Pete Torres, to network all the schools byinternal LANs within the separate school building, and provide, asfar as its $400,000 technology budget would allow, Internetconnectivity. The funds came from a one-time state appropriation fortechnology for schools.
Before 1995, Belen was typical of schools with arising, if scattered, interest in telecommunications. There were 5voice-grade telephone-line equipped classrooms, with modem access tothe New Mexico- wide TechNet (which gives teachers individual accessto the Internet from anywhere in the state). A few teachers hadpersonal computers and modems at home, from which they could dialinto, without paying long distance charges, any commercial,university, or other modem-equipped network, inAlbuquerque.
But also typical of such schools, there were onlya few early adopters of telecommunications. A once-math, science, andcomputer teacher at the Belen High School, Greg Anderson(firstname.lastname@example.org),took the full time position of Technology Specialist for thedistrict. Later he recruited one of his former students - RobertFranklin (email@example.com),who was now in an undergraduate degree program in computer science atthe University of New Mexico, to assist him. These two comprise thetotal technical staff of the district.
The Economics of Wireless
Of the available $400,000, a sum of $300,000 wasallocated to network all the schools, none of which had a LAN, andintegrate them into the administrative services computers of thedistrict. Firewalls isolated the administrative computers from theeducational computers. There were two school libraries and resourceson LAN servers needed throughout the district.
While the district wanted to connect itsspread-out schools to each other and the Internet, they did not thinkit was economically feasible. A happenstance contact with a localresident familiar with spread spectrum radios (a Sandia Test Labsworker) informed them that they could use commercial versions of suchradios to link their schools. So they put out an RFP specifying theconnectivity they aspired to: linking all schools with each other,and to one school that would be the hub, and through it to theInternet through TechNet in Albuquerque. It was a bold move, and hadthe support of the school board, even though this was a technologylittle known in the area.
Belen received bids of $800,000 that included amicrowave wireless solution, $550,000 for a hybrid wired solution,and $300,000 from an Albuquerque firm, Tamsco, which offered ano-license wireless solution.
Tamsco's quote used the Solectek AIRLAN/BridgePlus, which was rated at 1.544 Mbps, or T1 wireless operating inthe 915MHz range. Costs were $5,900 per radio, $1,200 for antenna,and $5,600 for the siting, erection and installation of the antennatowers and cabling. The high school was made the hub.
Ultimate cost, installed, was roughly $12,000 perschool to get the radio network up and operating at T1 speeds to andbetween all 8 schools and the district headquarters.
Before installation was complete, Solectekincluded their newer radios that operate in the 2.4GHz no- licenserange, at the same price. This, of course, required separate antennasat the hub, since the radios do not interoperate. All radios wereinstalled only as bridges (not packaged with routers). In the end,five 2.4GHz and four 915MHz radio systems were installed, linking allthe schools, and forming the backbone of the network.
Separately, three 3Com routers and a Cisco routerwere installed to route the TCP/IP packet traffic between sites, andto provide firewalls between the educational and administrativecomputers. 10Base-T Ethernet wiring was used throughout, supplementedby fiber links within some schools. Belen purchased 120IBM-compatible 486s, running Windows for Workgroups, and installedthem in the schools, linking them to the LANs and wireless netsthrough a combination of Novell NetWare 3.1, Windows NT 3.51, andLinux UNIX systems as LAN servers. The UNIX system provided TCP/IPname server and other network services at the hub site. This includedcentralized monitoring, and filtered if needed, access to the WorldWide Web or any Internet site from any school computer.
Essentially, with the T1 wirelessnetwork linking all the schools,
all workstations became part of a WAN,
using TCP/IP as the protocol.
Because the closest Point of Presence (POP) forInternet connectivity is in Albuquerque, 30 miles away, and theradios, using no more than the FCC Part 15-specified one watt, cannotreach that distance, Belen was compelled to look to their local phonecompany (US West) for dedicated line access to the city. The telcoquoted $1,000 to $1,400 a month (or $12,000 to $16,800 a year) for aT1 link to TechNet's POP in Albuquerque. So the district started withonly a 56kbps frame relay connection, using a $500 DSU/CSU, for whichservice costs $125 a month, or $1,500 a year. With similar quotesfrom the telco for local TI loop connections between the schools,comparative costs originally faced by the Belen District broke out atfollows:
By using wireless, the district could get T1 linksthey could not afford otherwise. The projected savings compared withthe only feasible alternative of telephone company-operated wirednetworks, was $320,000 over the first 5 years and $740,000 over 10years. These are very substantial economies for a smalldistrict.
This capability would not exist if not for theno-license radios, for reasons of recurring cost. There would be --according to Greg Anderson, the person most responsible for thesuccessful network -- no more than 8 separate local-school LANsinstalled and possibly, a limited 56kbps wired access from eachschool to the Internet. There would be no cross-district LANcapability, much less the multimedia (voice, sound, animated image,and text) interactive work that full education demands, and which thewireless T1 capability now gives.
Installation was done in the summer and fall of1995. By January, 1996, the wide area, Internet- connected networkwas in use by teachers. For at least one month after installation,there were severe reliability problems caused by software nothandling IP packets properly in the radios. These were corrected bynew software by Solectek.
Another difficulty stemmed from problems with the120-degree arc limitation of directional antennas Tamsco chose to useon the high school hub, while the three furthermost schools are in a150-degree arc from the point of the hub. They were not able to getan omni antenna to reach the distance required. So they were forcedto install two directional antennas on the high school hub's towerand relay through one school's radio -- feasible, buttricky.
When I visited the school district with DewayneHenricks, my Co-Investigator on this NSF Wireless Field Test Project,he noted other technical solutions existed. But we both agreed thatthere is relatively little wireless expertise in circulation yet.Until schools and colleges and the vendors who support them learnthese new fields, just as in the past they had to mastermicrocomputers and wired local area networks, there will be manymistakes.
We felt obliged to warn them that, under currentPart 15 FCC rules, they might be interfered with in both the 915MHzand 2.4GHz bands by both unlicensed and licensed devices, with norecourse. There was no particular protection for them, even though,to us, they were performing a very important public service withtheir radios. They were unaware of all the regulation involved, andonly dimly understood the technical issues.
But the network is in and it works quite well,with only minor problems its first full semester ofoperation.
As measured by ftp transfers of data, rates of upto 1.22 Mbps are achieved by the staff. I observed a transfer from anelementary classroom, by ftp, involving one wireless link and goingthrough routers to the Linux machine, of .997 Mbps, against the rated1.544 Mbps (T1) -- quite satisfactory. I also observed transfers andNetscape sessions through the wireless networks and the 56K link tothe Internet, in which the slowest link turns out to be the$125-a-month frame relay service. The fastest link is with theno-cost wireless.
With teachers only beginning to get used to theInternet resources and possibilities, the 56K link is adequate fornow, but Belen's technical staff is fully aware that T1 access to theInternet will be needed soon. The $12,000-a-year cost is a barrier,however. If they could use no-license wireless all the way toAlbuquerque, the district would be in excellent shape.
Teacher Use and Reaction
Teachers like the network. Some are unaware thatit is wireless, but all whom I interviewed - - the most experiencednetworkers -- were greatly pleased. They use it down to the lowergrades to access the Internet, as well as to reach centralized filesand programs on servers in schools miles away. The system, acting asone big LAN, works very well.
Both Jennifer Danner (firstname.lastname@example.org),a 4th-grade teacher, and Kassandra Boyd (email@example.com)a-6th grade teacher, demonstrated the network's uses in theclassroom. They showed their mastery of its commands by the speed,confidence and imagination of their applications. That, together withthe performance of the systems they were able to rely on, showedclearly to me that the wireless T1 network will be heavily usedthroughout the district, at all levels, for substantive educationalpurposes. Kassandra Boyd made one telling observation, when she said"My 6th graders don't like to use our 14.4 modem any more -- tooslow."
The district has practical plans to develop thenetwork further and also showed considerable interest in otherno-license wireless solutions to their problems of access. Forexample, there was interest in shorter- range wireless modems if theycould reach across the community of Belen (some 5 miles from thecenter to the district boundaries) from the homes of students andteachers. That would obviate the need for ever-larger banks of phonesto handle increasing study-from-home traffic.
They were even more interested in the possibilityof reaching an Internet POP in Albuquerque at T1 speeds bylater-model radios (some have a rated range of 25 miles for 2Mbs),but that would require a relay site at some intermediate site. Such arelay would cut the throughput to 1Mbps, 33% below full T1. But withUS West local loop T1 costs of $12,000 a year, or $60,000 for theusual 5-year contract, the cost of from $4,000 (bridge only) to$8,000 (bridge plus router) times three looks fairly attractive. Ifsuch radios, with better FCC rules, had a larger market, the pricewould come down.
The Superintendent of Belen schools indicated hewould like to see no-license radios with a range of 45 miles, so thatthey could reach a technical college south of Belen, as well asAlbuquerque, 30 miles away. He is expected to express that "need'" bye-mail to the FCC, which is considering NPRM Docket 102-93, theNII/SUPERNET device rules for "community network"-capable radios.Since Belen schools are used all summer for continuing adult, as wellas K-12 student education, the economics of the wireless, no-costdata links holds the glimmer of the possibility for broader"community networking."
Facilitated by their affordable, one-time cost,the no-license, T1 wireless radios have been the decisive ingredientin Belen School District achieving a district-wide network withspeedy access to the Internet from all schools. It is laying down thefoundation for advanced use of networks at every level at theschools, and is the beginning of community-wide access to both theirschools and the Internet.
As detailed in my paper "The Case for SharedWireless Spectrum" (May 6th, 1996), I have postulated that the84,000 public schools in 16,000 school districts in the U.S. havefour basic needs for no-license (or no-recurring-cost)wireless:
- Internal, to school buildings, wireless LANs. This is the least needed.
- Wireless high-data-rate links between the buildings of a school district. The greatest need.
- Wireless high-data-rate links between the central hub of a school district and the closest POP. A great need in rural towns.
- Wireless links, of at least 56Kbps, between students and teachers at home and their school -- as an access point to the Internet as well as access to school resources and communications.
The Belen School District, in its fulfilled needsas well as its unfulfilled needs, completely matches that model ofneeds. The FCC's rules for wireless should take those needs intoaccount and satisfy them.
David Hughes is a Principal Investigator forthe NSF. The Wireless Field Test for Education Project comprises twoyears of field tests on spread spectrum radios ("no license") inschools. E-mail: firstname.lastname@example.org
This article originally appeared in the 01/01/1997 issue of THE Journal.