THEnet FYI 01.1, Revised: April, 1994
PARTICIPATION IN THEnet
The Texas Higher Education Network is operated by the University of Texas System Office of Telecommunication Services (UT System OTS). The primary mission of UT System OTS via THEnet is to provide University of Texas component institutions with wide-area interconnectivity. This service includes access to the UT Austin High Performance Computing Facility, and to major Internet backbone networks.
As an important part of its mission, UT System OTS extends wide-area interconnectivity via THEnet to K-12 schools and school districts in the State.
A connection to THEnet is not a universally available service, but rather is the result of a limited cooperative arrangement. A school or school district qualifies to participate as follows:
1) through its status as a K-12 educational institution, and
2) through its willingness to undertake the technical, financial, and user support responsibilities necessary to maintain a viable, independent presence on THEnet.
Organizations must make a formal request to begin the connection process. The final section of this document contains a sequential checklist of the steps to be taken in the connection process. Staff from schools andschool districts are cautioned to read and thoroughly digest this document in its entirety.
All requests for participation in and connection to the Texas Higher Education Network must be submitted in writing to
William C. Bard, Director
A copy of the letter must be directed to:
Connie Stout, Director
In their letters of request, it would be helpful for educational institutions to explain the types of network services (e.g. electronic mail, file transfer, remote computation) to be used. An indication of the organizations (or types of organizations) with which connectivity is desired would also be helpful.
Certain administrative and contact information must accompany the request letter (see next section, Information to Accompany Request Letter).
INFORMATION TO ACCOMPANY THE REQUEST LETTER
The information indicated below must accompany the letter of request to the UT System OTS Director, preferably on separate sheets attached to the letter.
FAX:
For a direct connection, a school or school district must pay a connection fee of $100 per month. A school or school district contemplating multiple connections or aggregate bandwidth in excess of 1.536 megabits per second (T1) will be required to pay a higher fee. The connection fee is subject to occasional upward revision.
All schools and school districts must be willing to cooperate with UT System OTS to the extent that they will support their THEnet connection with local expertise, will lease a digital data circuit to connect to an appropriate THEnet point of presence, and will put in place and maintain at their own expense a local data communications infrastructure with equipment and software configured so as to be compatible with the state-wide THEnet infrastructure. Schools are urged to work with an equipment vendor or consultant who is knowledgable in the area of the Internet and wide-area TCP/IP internetworking.
The rest of this document treats the general requirements for equipment, infrastructure, technical registrations, and expertise for a direct connection to THEnet. Schools and school districts wishing to participate are urged to read this document carefully and to coordinate closely with UT System OTS.
UT System OTS must be notified AT THE BEGINNING of the planning process. No substantive plans should be made ahead of contacting UT System OTS. Contact with UT System OTS should be made through Internet electronic mail. Technical staff at schools and school districts should acquire an account that has Internet e-mail privileges, e.g. through TENET (Contact Connie Stout at 512-471-3241 or Richard LaGow at 512-463-9400 for details), or through an information service like Compuserve.
If no contact has yet been made at UT System OTS, then the initial contact may be made by sending e-mail to info@nic.the.net Technical staff at schools or school districts should become familiar with the concepts and applications of TCP/IP and the IP address space, including the Domain Name Service. Sources of information on these topics is included in Charles Spurgeon's NETWORK READING LIST, available in text or postscipt format, via ftp at ftp.utexas.edu in the /pub/netinfo/reading-list directory. Tech staff unfamiliar with ftp should consult a user-level book on Internet applications, e.g. The Whole Internet User Guide and Catalog by Ed Krol.
UT System OTS apologizes for the abruptness with which these requirements are made. However, knowledge of e-mail, telnet, ftp, and other Internet applications must be acquired sooner or later, clearly the sooner the better.
Before a direct connection can be made, one or more local area networks (generally Ethernet, Token Ring, or FDDI) must be operational at the school or school district's site. All users whose computers are connected to such a local network may gain access to the Internet and to some or all public information resources on the Internet.
By virtue of the THEnet connection, the computers on the local network will operate within a formally defined Internet domain. To support this, the school or school district must operate certain critical network services, namely Internet electronic mail and Domain Name Service. Domain Name Service provides the site with IP name-to-address resolution and an identifiable and reachable presence on the Internet. The school or school district is responsible for providing and maintaining the computer systems attached to its local network on which these critical services will run. Currently, the software necessary to implement these services runs only on computing platforms with multi-user, multitasking operating systems, such as UNIX, VMS, or VM.
Technical staff should then should consult with UT System OTS regarding the allocation of IP numbers for computers and other devices in their domain and the registration of a domain name.
The use of TCP/IP as the network/transport protocols for all desktop computers on local networks is recommended. Users of computers configured with TCP/IP may employ client software for electronic mail, telnet, ftp, gopher, WAIS, world-wide web, etc. This allows them to navigate the Internet, transfer files, and receive and send e-mail messages directly from their personal computers using software that has the look and feel of their desktop environment.
A direct connection requires that schools or school districts lease a digital data circuit for the carriage of data between itself and an appropriate THEnet point of presence, generally a University of Texas component institution.
The transmission capacity (bandwidth) of leased circuits is generally either 56 kilobits per second (referred to as 56K) or 1.536 megabits per second (T1). Such circuits are leased from local exchange carriers such as Southwestern Bell or GTE. Interexchange carrier service may be required if the circuit crosses LATA boundaries. The carrier will make monthly charges at tariffed rates and will also charge for installation of the circuit. THEnet staff can often provide helpful information concerning the lease of data circuits.
N O T E: A T1 circuit should have the following attributes: B8ZS data encoding, Extended Superframe (ESF) framing format, and line power. A 56K circuit has no optional attributes.
Schools and school districts must coordinate closely with UT System OTS regarding the installation of the data circuit. The circuit delivery date, order number, and circuit ID number must be transmitted to UT System OTS as soon as they are known. UT System OTS will coordinate with staff at the THEnet points of presence to insure that circuits are received and properly terminated.
Appropriate routing and line termination equipment must be purchased by schools and school districts, specifically a Cisco Systems router, a V.35 cable (which must be supplied by Cisco Systems due to the use of proprietary connectors on Cisco routers), and two Channel Service Unit/Data Service Units whose make and model are specified by UT System OTS (Cray 3180/3080 for 56K, Larse Access T for T1, see ahead in section CSU/DSU EQUIPMENT FOR DIRECT THENET CONNECTIONS for details).
On a campus or metropolitan network, or a wide-area network like THEnet, routers are responsible for choosing the pathway data packets will take to reach their destination. The CSU/DSUs are required to convert the V.35 synchronous serial data stream from the router into the digital signaling method used by the digital data circuit.
One port on the school or school district's router will connect to its local area network, e.g. into an Ethernet hub on the LAN. A serial port on the router will connect to the CSU/DSU via the V.35 cable. The CSU/DSU will then connect to the data circuit. On the other end of the data circuit, i.e. at the THEnet point of presence, the second CSU/DSU will link the circuit to the THEnet router there. The router and V.35 cable for the THEnet end will be supplied by UT System OTS.
If a school district intends to connect multiple separately routed campuses to THEnet, then the router at the ISD which 'gateways' to THEnet will need to have multiple serial ports, each leading through a CSU/DSU to a data circuit that in turn connects to a remote campus which is equipped with its own CSU/DSU, router, and local area network.
The choice of appropriate routing and CSU/DSU equipment involves several factors including manageability and expandability. These issues are discussed in the following sections.
Unless a Cisco Systems routers is used at the school or ISD, THEnet staff are unable to implement, manage, and troubleshoot the connection. This is so because important parts of the network management and diagnostic tools in use by THEnet engineering staff assume the presence of Cisco Systems routers on the THEnet backbone and at all school and school district connection points.
In cases where a school or school district's local area network is divided into separately routed networks and subnets, it is CRITICAL that the local routing equipment and protocols be chosen and configured so as to be capable of efficient, manageable, and reasonably error-free communications with the Cisco router at the THEnet gateway connection.
ROUTER PORTS, PORT EXPANDABILITY, AND 'DOWNSTREAM' CONNECTIONS
The school or school district's router connects to a CSU/DSU through a serial port.
The school or school district's local area network will connect to the router through an Ethernet, Token Ring, or FDDI port on the router. Multiple ports may be required to accommodate multiple local networks. Also, the presence of more than one type of LAN will require more than one type of port on the router (e.g. port w/ Ethernet interface and port w/ Token Ring interface).
A school or school district may agree to pass THEnet connectivity back to other local organizations in order to share the expense of the data circuit with the 'downstream' site. In such a case, the school or school district's 'upstream' router must have an additional serial port to connect the 'downstream' organization. A leased data circuit of appropriate bandwidth will run between the 'upstream' and 'downstream' sites with a CSU/DSU connected to each end, and a Cisco router at the 'downstream' site.
IMPORTANT: Before an organization connects 'downstream' to another, THEnet engineering and administrative personnel must be formally notified and must approve the connection. THEnet permits such connections with strong reservations that stem from the fact that the deeper the network becomes, the more difficult it is to manage and troubleshoot, the more extensive the effects of local outage, and the more likely it is that network packets will die through time-to-live decrements as the number of router hops increases. A 'downstream' organization must be aware that connectivity may be lost through an 'upstream' outage, and that resolution of such outage may depend heavily on the ability of the 'upstream' organization to authorize and follow through on diagnosis and repair measures. Both 'downstream' and 'upstream' organizations must understand that they are in competition for use of the available bandwidth, with diminished throughput a likely result.
NOTE It is THEnet policy that UT System OTS have full password access to all organization's Cisco routers at their external connection points. Organizations (including schools and school districts) are expected either to share operational passwords with UT System OTS, or at minimum to arrange the use of temporary maintenance passwords with adequate assurance that communication and cooperation can be obtained when needed.
CISCO SYSTEMS ROUTERS
For their THEnet connection point, schools and school districts will require a Cisco router that has at minimum a serial port for the 56k or T1 data circuit, and at least one local area network interface, generally Ethernet. As noted above, multiple local area network ports may be required.
For smaller sites, these needs are generally satisfied by acquiring a router in the Cisco Systems 2000, 3000, or 4000 series. The 4000 series offers some choice in the interface type and greater expandability in the number of ports.
For larger sites with numerous local networks and subnets, Cisco Systems AGS, MGS, or series 7000 routers may be advisable. Schools and school districts are urged to work with their equipments vendors or consultants in making this choice.
CSU/DSU EQUIPMENT FOR DIRECT THENET CONNECTIONS
As indicated above, a CSU/DSU is required for each end of the leased data circuit that connects a school or school district to THEnet. Schools and school districts must purchase a chassis- mounted CSU/DSU to be installed on the THEnet side and will generally want a stand-alone CSU/DSU for their own side.
The school or school district is responsible for delivering the CSU/DSU to appropriate staff at the THEnet point of presence.
IMPORTANT: For 56K connections, Cray 3180 (chassis-mount) and the Cray 3080 (standalone) CSU/DSU's are required. For T1 connections, Larse Access-T chassis-mount and standalone CSU/DSU's are required.
LOCAL EXPERTISE, MAINTENANCE, AND TROUBLESHOOTING
THEnet engineering staff will help with the initial configuration of the school or school district's router. Other technical functions and all user support must be assumed by the school or school district. The school or district should consult as needed with its equipment vendor regarding the installation and configuration of all hardware and software.
Router configuration deserves special mention. The school or district should work with its equipment vendor, as needed, to install and connect the router and the CSU/DSU, to test the data circuit, to place appropriate IP numbers on the router ports, and to assign appropriate passwords. The IP numbers and passwords will be assigned by UT System OTS. UT System OTS will then remotely configure the router.
Schools and districts are responsible for maintaining their router, CSU/DSU, and all local equipment, including all local network components, and all computers on the local network, including the host(s) on which electronic mail is received and domain name service is configured.
Schools and districts are responsible for acquiring, configuring, and maintaining all host- resident client and server software for electronic mail, domain name service, telnet, ftp, gopher, WAIS, World-wide web, etc.
Schools and districts are responsible for acquiring, configuring, and maintaining all software on personal computers and workstations, including communications software used for dial-up (Kermit, Procomm, etc.), and SLIP if it is used, as well as TCP/IP if it is used, and all client software that might be used on personal computers or workstations for electronic mail, telnet, ftp, gopher, WAIS, and the like.
Technical staff in schools and districts should be able to isolate and resolve all problems behind the router at their connection point to THEnet.
Technical staff in schools and districts should become familiar with the tools 'ping' and 'traceroute' to help assess the state of their THEnet connection. Technical staff should have at hand the IP names and numbers of local hosts and common destination hosts elsewhere on THEnet and beyond. They should know the IP names (if any), numbers and the passwords of their router. They should know the location of the CSU/DSU, and be familiar with its operation and display. In addition, they must be prepared to deal with local exchange and/or interexchange carriers for the resolution of any problems arising on the leased digital data circuit. Contact names, trouble numbers, and the circuit ID number should be readily available. In many cases, UT System OTS will coordinate the troubleshooting of the circuit and will communicate with the carrier.
In all cases where a problem is not demonstrably and solely locatable within the local network, THEnet engineering staff should be notified and provided with a narrative of observations made, when they were made, the diagnostic measures attempted (if any), and the results. Such notification should be made, if possible, by electronic mail (trouble@nic.the.net) or by telephone to the THEnet Network Operations Center (512-471-8530). THEnet engineering staff will help to isolate and identify the problem, and will delegate tasks to appropriate parties, including to itself, to the participating site, and to the local exchange and/or interexchange carrier. THEnet engineering staff may require any and all of the IP names and numbers, router and CSU/DSU information, and contact information for local exchange and interexchange carriers.
REQUEST FOR IP NETWORK NUMBER
An IP Network Number, or a CIDR-aligned block of network numbers from the CIDR-aligned portion of the IP address space allocated to THEnet, must be obtained by the school or school district. This must be done prior to configuration of domain name service and prior to the registration of domain name.
The template for requesting the IP network number follows below. Questions 1-2 and 5-6 should be completed in the template and the entire template returned by e-mail to UT System OTS. Questions 5-6 may require consultation with UT System OTS. Questions 3-4 are pre-answered.
IP NETWORK NUMBER REQUEST TEMPLATE [9/93]
(1) Provide the name, title, mailing address, phone number, and organization of the technical person of contact. The online mailbox and NIC Handle (if any) of the technical POC should also be included. This is the POC for resolving technical problems associated with the network and for updating information about the network. The technical POC may also be responsible for hosts attached to this network.
(2) Identify the the responsible organization establishing the network and the network geographic location.
(3) Networks are characterized as being either Education, Research, Defense, Government - Non Defense, or Commercial, and the network address space is shared between these five areas. Which type is this network?
(4) What is the purpose of the network?
(5) Supply the short mnemonic name of up to 12 characters for the network. Please use only uppercase letters, digits, or dashes ("-"). This name will not appear in the Domain Name System, but will appear in the InterNIC WHOIS database. As an example, if this network is for Fubar State University and this is the first network they have applied for, then FUBAR- STATE1 might be a reasonable name.
If you are requesting a block of network numbers, then each net number in the block must have its own unique name. Replicate line 5 below as needed to specify a name for each net number. Specify only one network name per line. For example, if you are requesting a block of four net numbers then you should specify four instances of line 5, each with one name on it.
(6) Specify the number of network numbers you need. This number must be a power of two. Each network number is suitable for use on one physical network (including extended networks connected by repeaters or MAC-level bridges which are IP-transparent), and may contain as many as 254 individual hosts. NOTE: While the CIDR specification does include the use of supernetting (that is, the assigning of multiple Class C network numbers to a single large physical network in order to allow for more than 254 hosts on that network), this capability is not supported in most current implementations of TCP/IP.
INFORMATION TO BE SUBMITTED BY E-MAIL AS SOON AS KNOWN
Schools and school districts connecting through THEnet will be registered in the isd.tenet.edu domain as
school-name.isd-name.isd.tenet.edu (e.g. lbj-high.austin.isd.tenet.edu)
or as
area-name.isd-name.isd.tenet.edu (e.g. admin.austin.isd.tenet.edu)
In general, domain names should be discussed with UT System OTS. The information template to be used for domain registration follows. Question 1 has been pre-answered. Questions 2-8 remain to be completed, 2, 7, and 8 in consultation with UT System OTS.
When complete, these should be submitted to UT System OTS.
DOMAIN-TEMPLATE [ 04/93 ]
(1)The name of the top-level domain to join (EDU, COM, MIL, GOV, NET, ORG).
(2)The name of the domain (up to 12 characters). This is the name that will be used in tables and lists associating the domain with the domain server addresses. [While, from a technical standpoint, domain names can be quite long we recommend the use of shorter, more user-friendly names.]
(3) The name and address of the organization establishing the domain.
(4) The date you expect the domain to be fully operational.
(5) The handle of the administrative head of the organization -- or this person's name, mailing address, phone number, organization, and network mailbox. This is the contact point for administrative and policy questions about the domain. In the case of a research project, this should be the principal investigator.
NOTE: Both the Administrative and the Technical/Zone contact of a domain MUST have a network mailbox, even if the mailbox is to be within the proposed domain.
Administrative Contact
(6) The handle of the technical contact for the domain -- or the person's name, mailing address, phone number, organization, and network mailbox. This is the contact point for problems concerning the domain or zone, as well as for updating information about the domain or zone.
Technical and Zone Contact
(7) Domains must provide at least two independent servers on Government-sponsored networks that provide the domain service for translating names to addresses for hosts in this domain.
Primary Server: HOSTNAME, NETADDRESS, HARDWARE, SOFTWARE
(8) The Secondary server information.
IN-ADDR REGISTRATION
The Internet uses a special domain to support gateway location and Internet address-to-host mapping. The intent of this domain is to provide a guaranteed method to perform host address to host name mapping, and to facilitate queries to locate all gateways on a particular network in the Internet.
The following information is needed:
IN-ADDR domains are represented using the network number in reverse. For example, network 123.456.78.0's IN-ADDR domain is represented as 78.456.123.IN-ADDR.ARPA
For example:
IN-ADDR domain Network Name IN-ADDR Servers
(Hostname)
(NetAddress)
(CPUType/OpSys)
41.192.IN-ADDR.ARPA NET-TEST-ONE zeus.central.lakeside.isd.tenet.edu 123.456.78.1
VAX-II/VMS
ns.abc.com
98.76.54.32
SUN/UNIX
The IN-ADDR template below must be filled out and submitted to UT System OTS. In general, the first IN-ADDR server will correspond to the Primary Server list in Question 7 of the Domain Template, while the second IN-ADDR server will correspond to the Secondary Server listed in Question 8 of the same template.
IN-ADDR-TEMPLATE [ 04/93 ]
Please have the Network Coordinator complete and return the following
information for those networks needing IN-ADDR registration.
IN-ADDR domain - - - Network Name - - - IN-ADDR Servers
CONNECTION CHECKLIST
[___] Submit request letter and accompanying information by US Post or FAX
[___] Obtain access to Internet e-mail
[___] Request IP network number and establish THEnet connection point and bandwidth via e-mail to UT System OTS
[___] Locate equipment vendor/supplier and draw up tentative purchase list
[___] Confirm items on purchase list with UT System OTS
[___] Arrange with local tech staff, vendor/supplier, or other consultant for equipment delivery, installation, and configuration
[___] Confirm timing of equipment delivery with UT System OTS
[___] Order data circuit from local exchange carrier to be delivered no sooner shortly before equipment delivery and configuration
[___] Confirm circuit delivery (order number, circuit id number, etc), and discuss desired date range for connection with UT System OTS
[___] After connection is made and Unix host is configured, submit domain registration templates to UT System OTS