Thursday, 16 June 2011

Adventures in a fibre database


I seem to have a habit of walking into jobs where there has been little to no documentation left by the previous inhabitants of the posts. Currently I am getting to grips with the fibre records for our institution, some of which it appears have not been updated since 1996.

Originally I was very excited to be getting more responsibility, however that excitement quickly turned into a realisation as to why nobody else wanted to do it. There are approximately 460 fibre trays around campus (that there is a record of - more on this in a bit) and hardly any have an accurate record of what is connected to them or whether they work.

It started as part of our resilience project, testing what links were usable and which needed re-splicing or more fibre coming in. It soon became apparent that there were various 'issues' with the records that were being kept:

1. Some trays marked with department names (rather than address or building code), and the departments in question have not been housed in those buildings for years. So unless you have been here 10 or 20 years you could spend all day/week/month looking for a link that does not exist.

2. Some trays had no labels on at all. Again this involved relying on the knowledge of people who could remember them being installed.

3. Some trays have one building name, some have another and some have a third... all meaning the same place.

4. Trays that don't exist (or phantom trays). This could take two forms. A) We have a record of it existing and having links on but following those links leads to nothing at the other end. B) The tray exists but there is no record of it, where it goes, what it is used for.

5. A tray is full of links, all or most are not labelled, the database says they are all free.

6. The database says the tray is all used, but when finding the tray it is empty or nearly empty.

7. The tray is either all single-mode or all multi-mode or split between the two but this information is not accurate on the record kept.

I have tried to impress upon people in the team the value of keeping the documentation up to date but largely seems to have fallen on deaf ears. As I found out the other day when, as an attempt to provide two links to a computer room to be used for registration (very important, don't want more cock-ups like last year) led to a whole load of wasted time.

It didn't help that the room was one of the furthest away from our two 6509's, but according to our records there should have been a route and enough spare fibres to do it. It just involved a long walk around the city.

After configuring the port-channels, patching and labelling I managed to get from Router to tray1, to tray 2, to tray 3, to tray 4 and from RouterB to tray 1b, to tray 2b, to tray3b before I arrived at a bottleneck where there were not enough spare fibres to continue (despite them being free on our database). Also on the database somebody put in that the fibre is 12+12 (12 single-mode and 12 multi-mode fibres) without actually checking. So I patched all single-mode and get to a tray which only has multi-mode on it and was led to believe there were 24 connectors when there were 8 or 6.

This added to the fact that many areas now have asbestos issues so we cannot go to a cabinet to check feeds from trays. So when we are asked by people who want to use our fibre (CCTV for example) 'what pair is this?', 'is it multi-mode or single-mode?' and the records are not up to date (or just plane wrong) we may have to get risk assessments and asbestos specialists to go and have a look/patch for us.

As a result I have started putting the record numbers on the trays (a unique ID from the fibre database). This should make the records a lot easier to find, rather than having to know what a building or department was called in the late 1990's. As I have been going round I have been trying to trace all the fibre in use and update that information. I have also had a look for any problems or issues that have or may arise, such as full cabinets, faulty fibre or damage.

Hopefully this will eventually lead to less leg-work walking miles to get some basic information so more time can be spent fixing faults and answering the helpdesk calls.

Our campus is rather large and with hundreds of trays to do this could take some time.

No comments: