Hi DN, I guess I'll never learn that I should read all of the posts, first, prior to putting up my own.
Please read my reply #3212 to Barry, and accept my apology if you thought I was discrediting you in any way. As it turns out, you do have more direct knowledge on this subject than I would have imagined from your brief posting on the subject of submarine cable placement, upstream.
=======
Re: sea floor placement vs suspension of cables, see the opening comments which were written in this highly relevant, yet dated, release from AT&T going back to 1992, which addresses this and other downstream comments and considerations which I've taken the time to review in advance (see? I'm learning), w.r.t. troubleshooting techniques for underwater cable breaks. I've provided a brief snippet here, followed by the url. Note that while the sonar device covered in the article was new at the time, it has been honed and perfected, and in some ways supplanted, by other techniques. But it is relevant to this discussion, nonetheless.
"AT&T today announced it has successfully tested a prototype digital sonar device that can find the newest fiber optic cable, plastic, metal, ropes and other materials suspended in mid-ocean or buried in seabeds.
"Existing ocean-bottom scanners require a magnetic field to locate coaxial cables and metals on the ocean floor.
"We've developed this device specifically to find today's small-diameter fiber optic cables that are buried at depths exceeding one meter beneath the ocean bottom," said Jim Barrett, vice president-engineering and operations for AT&T Submarine Systems, Inc., a strategic business unit of the global communications and computer company."
Continued at att.com
I too have had experience with snafus where cable cuts beneath the sea bed were entirely intractable and untraceable because of debris and mud flailing up as divers used blowers to uncover its ends after a break, necessitating long periods in between "blows" to allow visibility to clear. I've posted about this here in the past. It took place off of New England after the placement of TAT-5, the first time they used "sea plows" to bury submarine cables beneath the seabed on the continental shelves, so as to ensure that they were out of harm's way!
I am interested in finding out more on both the suspension aspect in the mountain ranges, and the remote troubleshooting techniques used, myself. So, if you or anyone else looking in here has links to those subjects, please post them and they will be greatly appreciated by me and others, I'm sure.
========
As for "self healing" in transoceanic rings, that's another matter entirely, and could consume many hours in discussion. Your depiction of how the "self heal" takes place in a classical sense, albeit a generic one, is accurate, and the same could and likely does hold true, from the standpoint of what is possible, for oceanic crossings as well. That is, with the noted extra time (hundreds of milliseconds vs the standard 50 ms) for restoral times, as you duly note, due to the added propagation times needed to cover the distances involved.
But experience and some forethought on the matter suggest to me that this may not be the case for all bundles of bandwidth that traverse the deep. In addition to the reasons I've already stated upstream, there may be other strategic (from a network philosophy perspective) and economic reasons, why such may not be the case.
But before I get to those, let me just say that some bandwidth, after it has been demuxed and electrified into manageable digital cross connect chunks, may be switched to the other arc (the other cable) of the ring by means of re-mapping, or by automatic protection switching (ASP), or by sparing and failover (M x N) techniques. Still others might be done by manual patching techniques.
And some bundles, like you suggest, could very well be administered under SONET/SDH headers in the classical sense, where wrap-arounds take place.
The specific techniques used could easily wind up being a matter of subjective, and many "business case" decision processes, on the basis of customer mix and what their requirements are, in combination with other factors which I've already stated, and those which I am about to state. For, there are more criteria to contend with, partially deriving from the latter, and some of which go back to what I was saying in my preceding posts.
Let me cover a few of them from a stream of consciousness, as follows.
1. Take the case where a customer is riding over AC-1 as a means of backing up their TAT-13 traffic. They are "already" backed up, in a sense, albeit they may only be doing load sharing. But the point here is, Why would they want to pay for tertiary provisions, if the laws of probability hold that they are statistically safe by going with two paths already? At the prices we're talking about here, you can be sure that they don't want to pay for a third or fourth pipe if they don't have to, especially when you consider the grief involved with dealing with some of the PTTs abroad (and I don't care who is aligning themselves with whom, on this matter) for getting just one link established.
2. Secondly, the transoceanics themselves are now boasting to be able to handle all IP over whatever, as a means of reducing dependency on SONET, if I'm reading them right. This plays into another related consideration, and that is,
"Which layer of the OSI RM do you want to do your restoration at?"
This is not an uncommon dilemma being faced by networkologists these days, and it's significance as a bugaboo is growing each day in IP backbone circles and in those situations which include multiple higher layer partnerships and third party service provisions, such as those which involve ASPs and bypasses.
Do you want to depend on Layer 1 (SONET/SDH) alone? Some would argue that this is unnecessary, and possibly redundant (pardon the pun), if a capacity glut is actually, and indeed, upon us. Just switch to another optical path and be done with it, and ignore the wraparound folly which often leads to hystrionics and recriminations by those who get "bumped" or preempted.
Do you want to restore at Layer 2, which would encompass SONET and ATM? AT what?!, yet others would ask?
OR, do you now look to Layer 3, where IP smarts begin to take over?
The point becomes somewhat moot when you consider that "ALL IP" means the absence of SONET/SDH, when taken to extremes, or when using dumbed down versions of these two workhorse technologies (SONET/SDH), as we stated would be the case in the Nexabit discussions here, recently.
3. And thirdly, there are the matters of grooming and fill rates on each arc of the ring (on each crossing, in effect) and how much -what number of channels- of the remaining capacities on each could possibly be reserved in an economical and prudent manner in "standby" status.
If each of the pipes (that is, AC-1 and AC-2) is being filled to beyond the half way mark, as I suspect will occur, this will actually very good from a revenue perspective. But it hardly supports the concept of total restoral capability, when you think about it. For, three quarters plus three quarters does not equal one. Something has to go over the side.
But does this matter when you consider the other factors I've stated above? Of course not. And such will be the case, and for all of the reasons I've stated, and some which I'd need more time to think about, I'm sure.
Anyway, it's been good chatting, and it's nice to know that you and others here are knowledgeable in this elusive art form. If you have any further data on the mid ocean placement methods now being used, or on troubleshooting techniques being used, please post the links here. TIA.
Regards, Frank Coluccio
|
