Re: Tkined - A few questions

Juergen Schoenwaelder (schoenw@gaertner.de)
Sun, 8 Jun 1997 16:35:21 +0200

David Coyle <dlcoyle@vivid.newbridge.com> said:

David> Bones: I was wondering about the Bones integration. There is not much
David> about this in the newsgroup and I see that it is not really integrated
David> in the sense of a direct database backend to the process of: Discover
David> and/or Data Import; Draw the map; Add/merge attributes; Save to
David> database. Question: Is tighter database integration forseen in the new
David> version or is someone working on such a project? I am new to TCL so
David> maybe I just don't "see" it. I suppose you could just do :
David> foreach object [inet retrieve] {decompose_and_insert $object $database}

Bones was a nice project where we made several experiments and learned
a lot about how to do it wrong. Database integration is really
difficult, at least if you do not want to dictate a particular
database system and a particular database model. I am currently not
working on database integration. Instead, I am adding the core of
Tkined into the Tnm extension which will lead to a "standard" file
format to describe network maps. This file format can easily be read
by every tool which is based on the Tnm extension. This will allow to
move to smaller, specialized tools that are able to use a common
network description.

Note that you can easily put these network maps into a revision
control system like RCS. It turns out that tracking changes is usually
very important and revision control systems are IMHO much easier to
use for this purpose than databases. I have also added a simple
mechanism which allows to store information about events and/or
statistical data in flat files. I think that this is sufficient for
many environments. Note also that there are very nice interfaces to
relational database systems for Tcl. It should be quite easy to write
a Tcl script which moves information to a database engine and back.

David> Telnet/Expect: I have been informed that security-conscious
David> customers are said to only allow "telnet" protocol through
David> (via dialup or the net) for troubleshooting purposes. If this
David> is a common scenario (is it?), then it is unfortunate to have
David> to forego every protocol except for telnet. So, question: has
David> anyone developed scripts to drive the performance
David> meters/bar-charts through tcl/expect/telnet processing
David> (instead of rpc or SNMP)?

Telnet alone is not secure. You have to run something like secure
shell (ssh) or a telnet session on top of the secure socket layer in
order to get security.

David> Remote: How does the Remote/Tkined work (in manager.tcl-
David> ComAccept/Listen/Connect) ? Does this mean that you can run
David> Tkined as a remote agent or server through rpc calls across
David> the network? I did'nt see much about it in the docs.

There was a remote interface several releases ago which allowed to
connect a local Tkined to a remote Tkined. This remote interface was
'proof of concept' study and had lots of problems. One reasons to put
the core of Tkined into the Tnm extension itself is to split Tkined
into a client/server application, where the user-interface is
separated from the management system itself. This will give us remote
access for free and it will allow to add new interfaces, like a Web
frontend. I am moving into this direction, but it will take some
releases until this is all in place.
Juergen

-- 
Juergen Schoenwaelder     <schoenw@gaertner.de>     (Tel: +49-531-23873-0)
Gaertner Datensysteme, Hamburger Strasse 273a, 38114 Braunschweig, Germany
--
!! This message is brought to you via the `tkined & scotty' mailing list.
!! Please do not reply to this message to unsubscribe. To subscribe or
!! unsubscribe, send a mail message to <tkined-request@ibr.cs.tu-bs.de>.
!! See http://wwwsnmp.cs.utwente.nl/~schoenw/scotty/ for more information.