Here's another take on telcom switches, in response to an earlier thread that may or may not have started with a "troll post" (this stuff..it's beyond me..).

I wrote a good friend, a technology expert recently retired from one of the Baby Bell's, to comment on the earlier post. He's a straight-ahead, trustworthy guy. His response follows. This, however, is not my area of expertise, so I can offer no further comment on it.

---------- Most modern switches run unattended 7x24 for 30 days at a crack. Unless there is a problem, someone visits the CO only once every 30 days for general checks. No LEC or CLEC staffs most COs. The writer is correct in a couple of areas. The cost of upgrading is ridiculous, though the path is clear. Some small LECs turned their clocks ahead (on traffic carrying switches no less) and found they worked just fine. The areas that are Y2K sensitive in the network are the billing systems. You or I might get a very confused bill, but we will have service on January 1.

Also contrary to popular belief, switches won't crash if everyone takes their receiver off the hook at midnight. The switch will give slow dial tone or fast reorder busy (fast busy) but won't crash.

Much of the equipment in switches is not date dependent therefore has noY2K problems. Most switches would switch traffic even if not remediated.

All the major telcos did inter-operability testing over the summer with advanced clocks. Some small telcos may go down for a short time, but they all will be staffed. The chances for long-term outages are less on January 1, then they are during the rest of the year. Everyone will be staffed up big time!

Bottom line MCI's Frame Relay Network went down for ten days this year and the world didn't end. MCI's failure didn't take anything else down. The chances of that kind of failure are higher during maintenance than during a "locked down" period like Jan. 1, 2000.

If used the author's method of reaching conclusions and applied it to cars, I never get in one! Traffic accidents kill 44,000 Americans every year. Each car carries at least ten gallons of gasoline. Gasoline can explode when ignited with the equivalent power of five sticks of Dupont 40% dynamite per gallon.

Hope this helps...

-- joe (joe@adeveloper.net), December 03, 1999

Answers

Most modern switches run unattended 7x24 for 30 days at a crack. Unless there is a problem, someone visits the CO only once every 30 days for general checks. No LEC or CLEC staffs most COs.

Most switches are "on premises" with other telco hardware at the CO where people are normally present. Switches are not always fiddled with but people walk by them all the time and switches do need regular tinkering with. A "switch" is not something you stick in a "closet" and bury. The reason they call them "Central Office" is because people usually fixate on them as the place where telco people congregate.

-- hamster (hamster@mycage.com), December 03, 1999.

joe said:
"Bottom line MCI's Frame Relay Network went down for ten days this year and the world didn't end. MCI's failure didn't take anything else down." hey joe-
It took down the Chicago Board of Trade. (during the outage)


-- plonk! (realaddress@hotmail.com), December 03, 1999.

regarding the "everybody on the line at the same time" -- a few years ago we had an earthquake in new england and everyone went for the phone at the same time in our small town. you know what happened? it was like a hotline. all of a sudden you could hear a zillion voices talking. strangest thing.

-- tt (cuddluppy@nowhere.com), December 03, 1999.

Joe is refering to the voice portion of central office equipment. Depending on the revision level of boards and software, there may or may not be any Y2K issues. Date related issues would come into play on system routing, and scheduled adds, changes and deletes for customers as well as any billing issues.

The Central office voice equipment is designed to be rock solid and most diagnostic and system management work takes place from a NOC (network operations center)

The data portion of central office gear is/can be a different issue in terms of 3rd party networking gear and other vendors equipment. This is more maintenance sensitive, requires a lot of time and attention in balancing and network engineering and management and could have more hardware related date issues. Not sure if MCI's frame relay problems were due to a Y2K problem but they sure created havoc for business customers.

Customer PBX equipment that is housed on site is connected to central office equipment via individual or grouped types of trunks. This on site equipment may have a date issue if there is time of day/day of week routing on long distance, historical stats and forecasting for call centers, voice mail, etc. All major vendors have published information on what is Y2K ready, what needs modification or what cannot be modified. (hence 50% of old 911 gear is not Y2K compliant-- a lot of that is real old equipment)

The biggest concern we all face with dial tone is:

panic level use such as you see after a flood or earthquake because the systems are not engineered to support every user picking up the phone at the same time

Multiple outages stretching the resources of tired people. It is a people and resource issue.

Long term electrical outages that exhaust robust battery back up systems.

Or some charming combination of all three...

With small, rural telephone coops or fringe level operating companies, the risks may well be higher. Overseas, who knows?

-- Nancy (wellsnl@hotmail.com), December 04, 1999.

It is odd that when I was working as a programmer in the telecommunications field that we downloaded the information off of the switch in order to bill the customers and every call had a date and time on it???? Where else do they think this info comes from???

-- Beckie (none@nowhere.com), December 04, 1999.

As an ex telephone company planner/engineer, I agree with Joe's comments concerning the basic voice network, However, the telephone companies have many support/maintenance/adminsitrative systems and most of them use dates extensively. I have no idea how much progress they have made in updateing these systems.

-- Dave (dannco@hotmail.com), December 04, 1999.