RIPE and European network resources registry

[Milan Sterba <]

At the last RIPE meeting in Amsterdam, we have extensively discussed
the need for a metainformation service about network resources and
accessible services at european level. The result of the discussion 
has been the founding of the NIDUS WG.

Network Metainformation Server
==============================

The general feeling was that the RIPE User Information Service (if defined
and set up) should act as a starting point for further search. It shouldn't
(in most areas) store the information itself but rather pointers to other
information sources and network resources and even pointers to pointers.

Lets call this function the Network Metainformation Server (because I want
it to cover a wider area than that covered by the current NIC's). 

1. In order to implement a NMS we need first to study the overall structure
   of the information and resources accessible today and in the near future 
   by the network and to define how the information in the NMS should be 
   ORGANIZED.

2. Then we have to define tools which would be suitable to support the
   type of data organization required (access tools, maintenance and update
   tools)

3. Then we must define the mechanisms (algorithms & responsabilities)
   to maintain the NMS up to date - links to other such activities on 
   the network (formal and automated)

At the last RIPE meeting in the NIDUS WG I have proposed to study 
a little bit what a structure of a RIPE metainformation server could
be. This has lead me directly to reflections about how could the
information about existing and future network resources be organized.

	Many things in many different ways are being done on many places in
	the area of network resource registry/discovery. One can easily be 
	wrong because of incompleteness, wrong tools used, insufficient
	availability etc.

Nevertheless a european resource registry is a reasonable need, the RIPE
community is probably one of the best places to start such an activity
and the RIPE NCC is a reasonable place to do it. That's why I propose
to use the well established RIPE database mechanism to register available
network resources in the european IP [CLNP] network. RIPE has succeeded
in implementing a pan-european data collection and maintenance machanism
which I find really suitable for the purpose of creating a pan-european
network resource registry.

Using the RIPE database framework would bring elements of solution
to all tasks 1.-3. stated above.

If there is enough interest in this topic I think it should be discussed
during the next RIPE meeting in Paris in the NIDUS and the DB WG's.
I will be happy to receive your comments before the meeting to be 
able to prepare a more detailed proposal (or to withdraw it if it doesn't
make sense).

Follows a proposal of a template of the new "resource" object of the
RIPE database. 
 
==================================================
Description of the "resource" RIPE database object
==================================================


resname:			Resource name in human
				Mandatory


admin-c:			Name of administrative contact person
				Format : <firstname> <initials> <lastname>
				Mandatory

tech-c:				Name of the technical contact person
				Format : <firstname> <initials> <lastname>
				Mandatory
restype:			Type of the resource
				Describes in one word the nature of the resource
				(the list of the most common should be agreed
				upon at RIPE)
				Format : one of the strings listed bellow

				meta 		= server of servers
				library 	= library catalogue
				docs 		= document store 
				database 	= structured data base
				local 		= local information system 
					(e.g. university information system)
				archive 	= specific data objects archive 
					(e.g. public domain software)
				super 		= supercomputer facility

				Mandatory								

acc-method:	 		Access protocol to the service 
				This should be the name of the IP [CLNP] 
				service which allows to access the server
				Format: sequence of bellow listed strings
					separated by space

				telnet | ftp | ftam | wais | gopher | www |
				e-mail | prospero | archie | ...
				
				Mandatory

acc-key:			Access info necessary to access the first
				level service

				(e.g. login name for public telnet sessions  or
				port number and database name for wais servers)

				Format: depends on the access protocol

acc-right:			Access rights to the resource
				Format : sequence of strings bellow separated 
					 by space

				ANY | see connect attribute in network object |
				LIMITED
				
				Mandatory

topics:				Keyword description of the topics covered
				by the resource
				Each topic should be mentioned in a new
				topic attribute.

				e.g. economy, mollecular biology,
				     computer networks, software

				Mandatory

description:			Short description of the resource
				Postal address is not needed, it can 
				be found via the contacts

host:				List of hosts for this resource
				Format: fully qualified  domainname
					without trailing "."

				Mandatory

media:				Type of data stored on the server

				Format: sequence of strings bellow
					separated by space

				ascii 		= ascii texts
				text		= other textual data
					(e.g. in native language or various
					text processing formats - PostScript)
				struct		= record oriented data
				graphics	= graphical data (images)
				sound		= self explanatory
				video		= dtto
				binaries	= computer programs in binary form  
				sources		= computer program sources
				etc.


==================================================================

Other open issues :

   - description of CLNP services
   - description of other services (accessible via other network protocols)







-- 

======================================================================
Prague School of Economics		e-mail : Milan.Sterba@localhost
Computing Center			tel : +42 2 21 25 704
nam. W. Churchilla 4			home: +42 2 823 78 59	
130 67 Praha 3				fax : +42 2 235 85 09
Czechoslovakia
=======================================================================