SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : The *NEW* Frank Coluccio Technology Forum

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Frank A. Coluccio who wrote (3687)8/14/2001 8:53:19 PM
From: TheStockFairy  Read Replies (2) of 46821
 
In this first post, I'm just going to give a little background information and delve a tiny bit into some of the mechanics of the process. At the bottom is what should drive the process to start from a customer standpoint. Some of this may be disjointed because I'm eating dinner :) I'll go through this again in a few hours to try and tidy up my answer. I'll probably outline in a very simple format what is needed for trading to really work, which is what, i think, is being asked here.

<<<with bandwidth on demand it is assumed that the bandwidth would be available for use immediately upon execution of a sale or trade>>>

Real time is well off into the future, although it could almost be done today with a good OSS system that will provision a circuit as a broker / sales rep is entering in the order. In some cases currently, you can actually get from a metro POP through the risers up to the customer suite, but that takes one hellova inventory tracking system. REAL firm order committment (FOC) dates are still pretty wishy washy, even in trading. If they weren't, why would there be liquidated damages? An actual order would look something like "if you want a circuit in Jan 02, tell me in Dec 01, if I blow the FOC date, Ill pay liquidated damages."

<<Alternatively, supply could be throttled dynamically, in an adaptive manner, through some parametric means embedded in software>>

You can almost do this now through a bandwidth manager like the Lucent Wavestar. You dump in a fiber connection to the bandwidth manager at both ends of the connection, your node would rest in the middle. There would still be some manual work that needed to be done at the software level, but it would eliminate a tech dispatch to LOC A and Z. Problem is, again, that a bandwidth manager to bandwidth manager connection isn't wanted by the carriers / clecs / enterprise customers, they want it to their own POP or suite.
Also, carriers aren't jumping at the opportunity just yet to lose control of their customer prem equipment, plus you can only connect to the customers connected to the bandwidth manager and you can't get building distribution rights to pull risers to the customers that are not connected.

<<Either way, it's bandwidth that is delivered rather swiftly on a demand basis.>>

See above, and add that currently a company would have to connect into the El Paso owned buildings, LighTrade Pooling Points, Poolingpoints.com, Equinix, Switch and Data and Laser sites to make this a good product. If you wanted to get really tricky, you would do all of the bulkheads and carrier POPs.

Keep in mind, none of this is centralized at this point, there is no Henry Hub like in gas trading. I don't think there is a clear understanding that not one carrier will have bandwidth all the time and have 100% connectivity to all the buildings, let alone interconnect agreements with all of the carriers and riser access into the buildings.

<<a bandwidth contract might stipulate a given pipe size or collection of pipes to wherever, which could be installed over a "normal" two-to-three month installation interval>>

all contracts must have a circuit speed associated with the order. you can accept shorter terms (monthly currently) by using carding, so you would install an OC-48 or 192 into several buildings, there you could put in tribs and break down to oc-12, oc-3 and ds-3 levels. reinstall a new system at 60-80% utilization. That would allow you to easily go on a month to month basis, using essentially the same circuit for different orders. If you accomplish this with carding, you cut your lead time down into 10 days and below install time, which is much less than the market is asking for now.

<bandwidth on demand>>

There isn't a ton of demand for this currently, I think we have gone over that subject before. Yes, there may be demand for it when it is available, but I'd rather make my money on the sure bets that don't tear your company apart by trying to change their entire business and operational rule set. Maybe new entrants whose sole focus is BOD will be better at supplying this service.

<<<But will it be - or is it, in fact, already - that simple?>>>>

Kinda, but mostly no due to the factors listed above. 50% is where you have connectivity to, 50% is a good OSS and inventory tracking system. In the market now, carriers only have portions of the above elements.

<<In even simpler terms, how does it work?>>

Ok, again, don't think in terms of BOD, it just isn't here and you are looking too far into the future, try to focus on the next two years. View the telecom market through Boeing's (for example) eyes.

Last year Boeing purchased a shiny new ds-3 going from one office in NY to another in LA the order will be placed on Jan 1. The sharp salesperson said "yes boeing our standard installation timeframe, with metro loops is 45 days (Feb 14). Boeing says "great!" and orders the circuit. The sales rep goes back in and enters an order for the DS-3 and puts in a requested due date of 45 days. The long haul provisioner gets the order, 7 days after it was entered (Jan 7), and laughs. He slaps a 45 BUSINESS DAY, not calander, timeframe on the long haul piece of the order, dated when he received the order (Mar 2). He then sends an order to two different local loop provisioners, one for Nynex, the other for PacBell.

The two provisioners don't go out and seek the best prices for the loop, they go straight to the LEC for the maximum price. The Nynex provisioner places an order with the LEC and they get a 30 day FOC on the install. The PacBell provisioner sends an order to PacBell and they get a 30 day FOC date.

Next, the long haul provisioner sends the order to the engineer who looks over his network. He looks to find capacity and it just isn't there. So he ends up routing the NY to LA circuit something like this: NY to Miami to Chicago, to Dallas, to Denver, to Seattle to LA. Lets just say for sake of argument that all equipment / fiber is there and ready to go. AND let's say this circuit tests OK.

Now, the long haul is in place, and I'll throw in that the Nynex loop was installed. On the day the order is supposed to be turned up, the customer sends his own install tec with a router to LA (pays for flight, hotel, food, car). The install tech is sitting in their suite waiting for the circuit that never shows up. The person who ordered the circuit calls the sales rep who calls his support person who calls the provisioner for the PacBell circuit. That person calls PacBell and they say "we are out of internodal capacity between Boeing and our node at that circuit speed, we need to do a fiber pull out to the site and install new equipment." Keep in mind, this is on the FOC date.

4 months later, PacBell finally gets their install completed, so the circuit that was ordered in Jan gets installed in July.

If anyone thinks I'm making this up.....think again, this is what happens every single day.

So in current market terms, what makes this work is more contractual than physical. If someone went to Boeing and said that they would deliver a circuit on the day they said they would, and if they didn't they would pay Boeing 125% of the of the pro rated portion of the contract for each day they miss, do you think Boeing would take that over best efforts?

<<If prearranged pipes that are sitting in the wait state for the purpose of connecting to a known carrier's cloud, most likely with an in-place set of SLA terms on a moment's notice, then these issues become less murky.>>

I'm not going to be the guy who signs the request form for circuits that are not going to be billing. You can do a few buildings on spec to get the sales rolling, but all new buildings have to go through an extensive senior management approval process before the expenditures are approved.

I'm not even going to go into how much it costs to light a building, including rent for the pop space, power, maintenance, building laterals, manhole installation, city franchise fees, riser access, COLOCATION COMPANY cross connect fees and port fees on the pooling point bandwidth managers.

Basically, let's not get ahead of ourselves worrying too much about semantics :). All poeple are requesting currently are circuits that are cheap and delivered on time. That's it. We will all worry about the trading aspect as soon as we are able to deliver circuits efficiently. (then the deployment will look like the dsl deployment where only 10% of the country can actually get traded loops to their suite, the rest will have to go through the RBOC with no SLAs.)

<I should be able to receive said bandwidth and connectivity from just about anyone in the pool at the lowest market price>

Which pool do you want sir? We have a choice of several, do you have connectivity to these pools or do you have a metro carrier in your building that has the interconnections in place to get to your pooling point of choice? If you only have one carrier that can deliver your circuit, is he going to be the cheapest market solution?
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext