Search This Blog

Friday, October 29, 2010

Prime ministerial

So it's the day after the final BEAR push, leaving me time to ponder questions such as, "What exactly is a 'ministerial error,' anyway?" So I looked it up and found this definition in the Code of Federal Regulations, C.F.R. 19 §351.224(f) :
Definition of “ministerial error.” Under this section,
ministerial error means an error in addition, subtraction,
or other arithmetic function, clerical error resulting from
inaccurate copying, duplication, or the like, and any other
similar type of unintentional error which the Secretary
considers ministerial.

(It's from Customs regulations, but that's close enough for government work.)

By the way, I checked, and Bishop Perry (the patron saint of "ministerial and clerical") does not define "ministerial." The FCC actually asked for a definition of "ministerial" in the 2005 NPRM, but they don't seem to have adopted one. I don't remember seeing a definition in any of the responses to that NPRM, but I definitely could have missed that.

Thursday, October 28, 2010

Smarter than the average BEAR

So here's a positive data point from this year's BEAR season: we only had to do three BEARs on paper. More and more service providers are getting the hang of the online certification over the past couple of years, but last year we still had to do a fair number on paper.

One nagging problem with online certification: I haven't seen the service provider interface for BEAR approval, but apparently the button they press says "Certify." So newbies take a look at the form and click on "Certify." But if they don't check a little box next to one of the FRNs, that FRN is cancelled. Most BEARs are only one FRN, so it's easy to miss that check box. So every year we get a handful of canceled BEARs, which we then have to redo. Even experienced certifiers forget to check the box sometimes.

There should be a warning when a service provider clicks the "Certify" button that says, "You didn't check the little box next to the following FRNs. Click 'Accept' to cancel those FRNs, or click 'Edit' to check the box before certifying."

And while I'm giving out BEAR beefs, why can't I modify a BEAR after submittal? It happens every year, more often than I'd like, that a service provider says, "It looks like you included this ineligible charge in your amount, fix it." Unfortunately, I can't fix it, so the SP has to cancel the BEAR and I have to post a new one.

Before a BEAR is certified by the SP, I should be able to make changes to it.

Monday, October 25, 2010

Ho Ho Ho

Mel Blackwell does not have a Saint Nick physique, but he certainly brightened my holidays.

USAC released a News Brief today (if I ran USAC, any News Brief that came out on Monday would start with "Extra! Extra! Read all about it!") which said:

  • We expect the FY2011 filing window to open in early January 2011.
  • The window will remain open for about the same number of days as in the past, which means that we expect the window to close in mid-March 2011.

I have been saying for years that the window should be later, and I finally get my wish. I'm sure it's only because USAC is scrambling to get the online 470 and 471 updated to meet new FCC requirements, but my hope is that after they try it for one year, it will become the norm.

Monday, October 04, 2010

E-Rate, not E-rate

OK, one final rant from the Sixth Report & Order.

Faithful readers will know that I have my knickers in a twist about the spelling and capitalization of "E-Rate." After belittling others for spelling it "eRate," someone pointed out to me that the FCC spells it "E-rate." I tried to discern Congressional intent, but as one might expect, that just muddied things further.

While reading the 6R&O (it's high time we had an acronym with an ampersand, don't you think?), I was captivated by the title "E-Rate Deployed Ubiquitously." "Deploy" and "ubiquitously" are just splendid words. As I savored the name, I realized that it proved that my capitalization was right all along.

See, the FCC capitalized the "R." Of course, they always capitalize the "R" in headings. Wait a minute. Why would it be capitalized in headings? Only the first letter in a word is capitalized in a heading. But the first letter of a word is also capitalized in a proper noun like "E-Rate." So the only logically consistent position is that the "R" is always capitalized. (Well, I suppose you could make the case that it's not a proper noun, but then in the middle of a sentence you'd get "e-rate" and that's just immoral, so let's not go there.)

Time to launch a campaign for an NPRM about the spelling of "E-Rate."

Sixth Report & Disorder

I am in full gripe mode this morning, so I thought I'd throw all my minor gripe about the Sixth Report & Order into one post. I've already posted my major gripes.

Again, I just want to make clear that as a whole, the 6th R&O improves the program. But the positive changes were all touted in the FCC press release, so I'll stick to the dark side.

  1. “As long as schools are in compliance with CIPA requirements, we leave specific policymaking decisions up to individual schools to address.” (footnote 76)
    So now that schools are offering Internet access to adult community members, do they now have to comply with the heinous CIPA rules that the Supreme Court came up with for libraries? That would mean turning off filtering for individuals who request unfiltered access….
  2. “we plan to include a box on the FCC Form 471 when we next revise this form for applicants to check if they are taking advantage of this rule change.” (paragraph 25)
    You know, I was just thinking that there aren’t enough checkboxes on the Form 471.
  3. “an eligible school or library may assess computer fees to help defray the cost of computers" (paragraph 26)
    Does a converse to the Free Services Advisory apply? Can my school charge $5/hour for computer use, and throw in Internet access for free, or will I have to do some cost allocation? Look for the blossoming of PTO-sponsored cybercafé fundraisers.
  4. “the cap for funding year 2010 will be increased to $2,270,250,000.” (paragraph 36)
    Well, at least it's in the right direction. At this rate, 40% schools will get some P2 funding around, say, Stardate 2542.7.
  5. “E-Rate Deployed Ubiquitously” (paragraphs 44-50)
    First off, kudos for using the word "ubiquitously." You just don't see that word often enough. But $10 million dollars to a “handful” of applicants who have already deployed off-campus wireless Internet access? Is there really a pool of schools and libraries that are spending $1 million/year or so on off-campus access? And who gets stuck evaluating the proposals?
  6. “we direct USAC to make available on its website and update on an ongoing basis a list of donation and recycling locations for communications equipment.” (paragraph 75)
    This is not how USAC should be spending USF funds. Remember the Eligible Services Database? I foresee cobwebs on this database.

Maintain this

What's the worst news from the Sixth Report and Order?

“We find that an unbundled warranty is an ineligible BMIC service because it is purchased as a type of retainer and not as an actual maintenance service.” (paragraph 106)

OK, first off, the reasoning is faulty. If an applicant pays every month for a phone line to connect an alarm system to the security company, that line is eligible, even if the alarm is never tripped. Inside wiring maintenance costs are eligible, even though wires don't wear out. To say that a warranty is not an actual service just because you don't use it every year doesn't hold water for me.

I understand where this is coming from. I, too, am appalled at the cost of manufacturers' warranties on phone systems and data network equipment. It definitely feels like they have us over a barrel. But it doesn't rise to the level of abuse of the program. The exorbitant amount that PBX manufacturers charge to schools and libraries is the same amount they charge everyone else. No one is gouging the E-rate.

What effect is this going to have? On the phone side, the answer is Centrex.

I have a meeting this week with a client that is about to sign a new 5-year contract on their phone system, at about a half million per year. Take a look at their options. It would be foolhardy to use a phone system without a service contract, so unless they can come up with a half a million a year out of their own pockets, they'll have to turn off their phone system (and since it was purchased 2 years ago with E-rate funds, let it sit in the corner for 3 more years until they can sell it for scrap). Now normally, they could get a nice deal on some hosted VoIP, and they might be able to cover the district for maybe $1 million per year, doubling their demand for funding. But now their data network won't have a warranty, either, so they can't jeopardize student safety by running their voice over that network. So they'll have to see if they can find a phone company that will give them Centrex service. With Centrex, the wiring will all be installed and maintained by the phone company, and the equipment will be back in the CO, maintained by the phone company. What will that cost? I'm estimating $3.5 million/year.

So the E-rate will pay an extra $3 million per year to move this district backwards to a technology that was cutting-edge in the '60s.

On the data side, applicants will probably just pay out of their own pockets for a spare edge switch, and swap it out when a switch malfunctions. Of course, the district will have to hold on to the malfunctioning switch for 5 years, because they really won't be able to replace malfunctioning hardware, but the switch probably still has some value, so they can't just throw it out, and E-rate rules won't let them sell it to someone who might fix it.

Core switches are more problematic. Applicants will be forced to design networks which can still function in the absence of any single core switch. Those networks will be more expensive, and will likely be limping along if a core switch does fail. More bad engineering made cost-effective by unfortunate E-rate rules.

There is another possible workaround, depending on how this rule is interpreted. It is clear the FCC wants applicants to set up maintenance contracts which set up a pool of hours, where the service provider is paid for hours actually used. If such a contract could also include a pool of money for purchasing replacements for failed components, then maybe we wouldn't see a resurgence of Centrex.

Whether the FCC allows applicants to throw replacement parts into the pool, setting up these pools means:
  1. Applicants will try to set up a pool big enough to cover the worst-case scenario, meaning lots of approved funding will go unused every year (which I don't see as a problem, but which has been called out by the GAO and Commissioner McDowell). And the maintenance cost-effectiveness tug-of-war, which already eats up too much PIA time, will require a whole separate division of Solix to manage.
  2. No matter how big the pool is allowed to get, the applicant runs the risk that there will not be enough funding to replace equipment wiped out by a major catastrophe. Just when the applicant needs help the most, it won't be there.
The pools will increase demand for funding, squeezing the fund in the short term. After a couple of years, the big pile of unused funds from the pools will start flowing back into the fund, and in the end I'd guess that disbursements will actually be lower than with warranties, except in years with large natural disasters.
While I agree that manufacturer warranties seem exorbitant, there is a reason that such warranties are "ordinarily provided in the marketplace to entities receiving such services without e-rate discounts."*
* Third Report & Order, paragraph 23, where the FCC defined Basic Maintenance.

My head is SPINning

Tucked away in the Sixth Report & Order is an offshore earthquake which will provoke a tsunami of lost funding and appeals to the FCC:

“[T]o alleviate uncertainty regarding the types of SPIN changes that are permissible following a competitive bidding process, we clarify that once a contract for products or services is signed by the applicant and service provider, the applicant may not change to a different service provider unless (1) there is a legitimate reason to change providers (e.g., breach of contract or the service provider is unable to perform); and (2) the newly selected service provider received the next highest point value in the original bid evaluation, assuming there was more than one bidder.” (paragraph 92)

That looks reasonable. And it addresses an area of actual abuse: service providers eschewing the competitive bidding process, then swooping in after funding approval and taking the business.

But most SPIN changes are not abuse. They are applicants who got a better deal on, say, long distance and want to switch providers. So now if a school district finds an opportunity to save money on their month-to-month phone service, they have to wait until they can file a Form 470, wait 28 days for no bids to arrive, then make the switch. For a lot of applicants, it won't be worth the hassle, so they'll forego the savings opportunity.

But that's not what worries me. I used to be a federal procurement officer, so I'm used to regulations that tie the hands of ethical employees and raise costs in order to make abuse more difficult for the unethical. The problem I have is when this rule hits the real world.

Because in the real world, the person who decides to switch phone carriers is often not the person who files for the E-rate. When I go to a client at the end of the funding year to collect bills to file a BEAR, I often hear, “Oh, we forgot to tell you: six months ago we found a cheaper phone company and switched service.” In that case, they will simply lose six months of funding. USAC won't be able to help them, so they will file a Request for Waiver with the FCC. It's going to happen a lot.

The only hope is that getting a lower price will be considered a "legitimate reason to change providers." Then it will just be a matter of hoping you only got one bid in response to your 470. However, if you got 2 or more bids, you are now locked in, no matter what kind of savings comes along.

SPIN changes were one tool to help applicants work within the broken E-rate procurement process to get the best deal. Now that tool has been ruined. Opportunities for cost savings will be lost, and innovation will be stifled.

Sorry, but I think the solution is worse than the problem.

Some shadows on dark fiber

I'm thinking I'll do a few posts about the Sixth Report & Order, because, well, it is HUGE.

I'll start with dark fiber. In general, it's great that dark fiber is eligible. And great that lit fiber can now be purchased from non-telecom providers. But of course, I'm going to gripe about some details. (Sometimes I think I should call myself the Angry E-Rate Nerd, but the Angry Video Game Nerd would probably sue me. Plus I can't seem to produce the requisite vulgarity.)
My gripes:
  1. “Although lit fiber is already eligible for funding as either a telecommunications service or an Internet access service … an applicant cannot lease the lit fiber for voice telecommunications from a non-telecommunications carrier.” (paragraph 11) I'm sure this isn't supposed to mean what it actually says. It seems to say that fiber leased from a non-ETP cannot be used to transport the district’s internal VoIP between locations. And that if a district has one of the new concurrent-call hosted VoIP services, where the telecom service is riding over the district’s WAN, the WAN must be leased from an ETP. I hope what is meant is that voice service has to be purchased from an ETP, regardless of what transport it rides over. I just wish that's what the order actually said.
  2. “Providing services using dark fiber may involve a number of additional costs beyond lease payments for fiber connectivity, and those costs should be factored in to a total-cost comparison across bids.” (paragraph 19) Oh, the pain. So let's say a typical client seeks a WAN, and they're trying to decide between a typical telco WAN offering, which would give them a live Ethernet port, or dark fiber, which would give them an unlit fiber pair. To use the dark fiber pair, they’ll need to put a fiber GBIC in their switch. The fiber GBIC is not eligible in a Priority 1 FRN, so it can't be included in the "primary factor" cost calculation. So now is the district required to create a separate criterion for total cost? Or one for ineligible costs? Either way, a disgruntled telco bidder can say it's not an "apples-to-apples" comparision, since the cost of the GBIC will be weighted less heavily. What about the cost of the GBIC slot in the switch? And what about the cost of the Ethernet port on the switch that the lit WAN would use? Do we include that cost? How do we calculate the cost of a single Ethernet port on a switch which has 24 copper ports, a fiber port, and a couple of GBIC slots? Now imagine the headache of calculating the cost of a GBIC+slot and Ethernet port on a chassis switch with a half-dozen blades in it. There is not enough caffeine in the world.
  3. “We include as eligible maintenance costs and installation charges.” (paragraph 19) “This includes charges for installation within the property line.” (footnote 52) “For purposes of the E-rate program, we will consider Indefeasible Rights of Use (IRU) purchase arrangements as a lease of dark fiber. To the extent an IRU contract contains significant upfront charges, and consistent with our existing requirements regarding upfront costs associated with the purchase of telecommunications services, applicants must amortize upfront, non-recurring charges where the upfront charges 'vastly exceed' the monthly recurring charges.” (footnote 51) Finally, some good news. It seems the “unused capacity” statements in the press release were more vision than rule. Go ahead and build a new dark fiber network, as long as the construction costs are amortized the way we currently do with lit fiber networks.
  4. “[W]e decline to extend support to cover special construction charges that may be incurred to build out connections from applicants’ facilities to an off-premises fiber network” (paragraph 19) “Special construction charges include costs for design and engineering, project management, digging trenches, and laying fiber.” (footnote 54) Wait.... What?! Did the FCC just say that when leasing fiber (lit or dark), the cost of running the fiber from the pole to the property line can’t be included? Really? Applicants can pay the service provider to run fiber all over town, but not for that one short segment? OK, but what is an applicant supposed to do in the real world? The typical fiber lease is a 40-year IRU, with the buildout cost amortized over the first five years. So the cost to the applicant is, let's say, $2,000 per month for the first 5 years, then $200 per month until the end of the contract. All buildout costs ("special construction" or otherwise) are included in the monthly cost. Is the service provider now required to dig into that cost and separate the cost of connecting the client to a pole from the cost of building out the rest of the WAN? If so, look for lots of contracts where the service provider actually runs their network into the client's building, so there is no special construction required. Another example of bad engineering made cost-effective by convoluted E-rate rules. Let's hope that what this really means is that if your contract has a separate line for special construction, you have to cost-allocate it out of your request.
Maybe I'm reading too much into the sentences I quoted, and my concerns are unwarranted. But I won't be advising any clients to get dark fiber this year. Let some other brave souls be the PIA beta testers on this one. I don't want to sign a 40-year contract based on Version 1.0 of the rules.