Search This Blog

Wednesday, January 22, 2020

Part-time student? You don't count.

If I'd gotten around to a full review of the FCC's recent C2 Order, I would have mentioned the problem with excluding part-time students, but Infinity Communications & Consulting did even better, submitting a Petition for Reconsideration.

Basically, the FCC went too far. A few commenters on the C2 NPRM pointed out that having district-wide budgets would simplify the application process for some districts, where students were part-time in two schools in the same district.  True enough.  But for some reason, the FCC went further and said, "In another effort to streamline both the application filing and review process, going forward we will base student counts on full-time enrollment only and eliminate the need for schools or school districts to count part-time students in their enrollment numbers."

I suppose that's fine in states where county-wide districts almost never send students to another district for part of the day.  But not here in NJ (or, apparently, in California, since that's where Infinity is based).  Here in NJ, there are districts that are mostly part-time students.  How?  Well, in NJ, we have more school districts than municipalities.  So way back when, the state created county-wide vo-tech districts.  The model was this: students take all their academic classes at their home high school, and then go to a vo-tech school for part of the day if they want to study car mechanics or cosmetology or culinary arts or whatever.  In NJ, they're called "shared-time" students.  So the vo-tech districts generally had few or no full-time students.  That's been changing at some of the vo-tech districts recently, as the vo-tech districts have expanded into full-time programs, but some of the vo-techs still have a significant portion of "shared-time" students.  Now, the state of NJ reports "full-time equivalent" enrollment numbers to USAC, so a shared-time students only counted as half a student.  That seems fair, no?

But under the new rules, shared-time students can't be counted by either district.

I probably shouldn't point this out, since USAC probably doesn't realize that the state has been including half-students in their enrollment numbers (though they should have realized it, since enrollment totals for a school can be, for example, 232.5).

So how about instead of "simplifying" things by not counting part-time students, the Commission just says, "Whatever the state says your enrollment is, that's your enrollment."

Miffed about NIFs

Once again, SECA is the voice of reason.  They've submitted a Petition for Reconsideration, asking the FCC not to implement the heinous new shared C2 NIF cost allocation rule.  What am I talking about?  It's a rule (tucked in paragraph 51 the new C2 Order) that requires applicants buying new shared C2 equipment (a firewall used by the entire district, for example) to reduce the eligible cost if that equipment serves a NIF in addition to schools.  So if you've got 7 schools and a board office sharing an Internet connection, and you need a new firewall, it's only seven-eighths eligible (unless you use a different cost allocation method).

The SECA letter makes a number of good points, but it left out 2 compelling examples of complexity: consortia and changes.  (I mentioned these issues back in 2014, when the C2 budgets first appeared.)  Since SECA has a bunch of statewide networks in it, it's surprising that it didn't highlight how difficult this will make life for large consortia.  Suddenly a consortium has to keep track of all its members' NIFs in order to know how much it needs to cost-allocate out of any new shared equipment purchase.

The changes problem is this: if a district adds a NIF to its WAN, that retroactively changes the eligibility of all shared equipment.  So if a district bought a firewall to cover its 7 schools last year, and next year decides to move the board office into its own building, suddenly that firewall is only seven-eighths eligible, so the district should self-COMAD and self-RIDF.

What really sets my teeth on edge?  In paragraph 51, the Commission says that they're not changing the rules.  Then on page 29, they change the rules.  The rules didn't used to say: "When applying for category two support for eligible services to a non-instructional school building or library administrative building, the applicant shall deduct the cost of the non-instructional building’s use of the category two services or equipment."  Now that is in the rules.

Monday, December 23, 2019

Pony up for the filters

I'm intending to give a more fulsome review of the new C2 Order, but today I'll single out one thing that seemed wrong: "the Commission has previously explained that the Children’s Internet Protection Act prohibits recipients from obtaining discounts under the universal service support mechanism for the purchase or acquisition of technology protection measures necessary for compliance with the Children’s Internet Protection Act."  Wait, what?

So I went back to the FCC's 2001 CIPA Order, and sure enough: "CIPA clearly prohibits recipients from obtaining discounts under the universal service support mechanism for the purchase or acquisition of technology protection measures necessary for CIPA compliance."

The FCC reached that conclusion based on § 1721(g) of the Act that created this mess, which says: "Notwithstanding any other provision of law, funds available under section 3134 or part A of title VI of the Elementary and Secondary Education Act of 1965, or under section 231 of the Library Services and Technology Act, may be used for the purchase or acquisition of technology protection measures that are necessary to meet the requirements of this title and the amendments made by this title. No other sources of funds for the purchase or acquisition of such measures are authorized by this title, or the amendments made by this title."

OK, I see that CIPA does not authorize the use of E-Rate funds for filtering.  But I don't think "not authorized" means the same thing as "prohibited."  It seems to me that Congress intended to ensure that applicants could use IDEA and LSTA funds for filtering, by overruling any provision that might have prohibited it ("notwithstanding") and to ensure that no one thought CIPA required other funding programs to change in order to allow funding of filters.

At one point, the FCC seemed to have some ambivalence on the subject.  The NPRM for the 2009 ESL wondered "whether [§ 1721(g)] explicitly prohibits E-rate program funding from being used for filtering software or whether the statute can be interpreted so that the Commission is not precluded from funding filtering software through the E-rate program." (paragraphs 14-15)  EdLiNC gave the most complete response, looking at legislative history and all.  Unfortunately, they came out against filtering.  Would EdLiNC reach a different conclusion now that we no longer hit the program cap?  (Back in 2008, any time a new service was included in the ESL, it meant the funding ran out earlier, and fewer applicants got Priority Two funding.)

So I disagree with the FCC.  CIPA does not "prohibit" the use of federal funds for filtering.  The law expressly allows the use of IDEA and LSTA funds (overruling any existing rules to the contrary), and makes no statement on whether other federal funds can be used.

I think filtering should be considered like any other service when it comes to E-Rate eligibility.  Which means it comes down to: are filters necessary to deliver broadband to classrooms and public areas of libraries?  My answer: "yes."  Or more precisely: "It's necessary if the applicant decides it is."  Almost all schools filter, because they know that the Internet has lots of content that is not conducive to education.  Lots of (most?) libraries, on the other hand, don't filter adult access, because they don't want to be in the position of censoring access.  (Unfiltered libraries can't get E-Rate funding for Internet access and Category 2, of course.)  I'd be interested to know what percentage of libraries allow young children to have unfiltered Internet access.

The other test I often think of is: "Do organizations which do not get E-Rate funding use the service?"  The answer is yes.  Well, at least it is here at On-Tech.  I make sure all our Internet access is filtered, just so we don't stumble on something unsavory.  It also protects us from malicious sites.

Let's add filtering to the ESL.

Tuesday, September 24, 2019

Talkin' EPC blues

I hear lots of discussion on C2 budgets and the merits (or demerits) of big consortia, but I haven't heard anyone discussing the change that affects me every day: the new EPC color scheme.

You noticed, right?  At first, I thought I'd messed with my screen brightness, but I confirmed that I wasn't the only one seeing it: the color on the EPC navigation bar at the top of the screen has changed.  For the old-school HTML geeks, it looks like it's gone from around #003366 to #0033CC.  Around here, it all looked purplier, but in fact, it's just bluer.

I don't like it.  It's too bright, and somehow the green made the page look more grown-up.  The new color is somehow cartoonish.  Yes, it's closer to the dark blue in the "paper flying at you" USAC logo, but I think the page looked better with more separation between those colors.

They also changed the text in the ... what shall we call it? ... navblob of links in the upper right corner below the navbar, and in other places.  The color is too bright for text.

Another minus: in the old color scheme, whatever was selected in the navbar had white text with a yellow underline.  Now the underline is white.  This doesn't make so big a difference, but I think the extra color was usefully eye-catching.

I can't really add this to the pile of reasons for scrapping EPC, but EPC is just a tiny bit worse now.

Tuesday, September 03, 2019

PIA's newest employee

Check out who is now doing PIA reviews: Commissioner O'Rielly!

This week, he sent a letter to the Superintendent of one the the consortia that he has been accusing of wasting money by overbuilding.

It's basically a special cost-effectiveness review (CER).  CERs from USAC are arbitrary and mysterious enough, but now we have an FCC commissioner dreaming up more requests. We've always been dealing with secret cost-effectiveness standards, but now we have a Commissioner creating his own standards after the application's in review.

Hey, here's an idea: before we start with multi-directional CERs, the Commission should first decide if CERs are allowable.  There are a large number of appeals of CER denials that have been awaiting FCC decision since 2008. Here's my dream: that in deciding those appeals,  the Commission sets clear standards for cost-effectiveness, which are then published every year in the Eligible Services List.

I've said more than once that I don't think that the FCC should be encouraging consortia, but it seems unfair and capricious for the Commission to be encouraging consortia consistently for 20 years, and then suddenly requiring an applicant to justify creation of a consortium and justify its members.  First, the Commission should clearly disavow paragraph 476 of the Universal Service Order: "we should encourage schools and libraries to aggregate their demand with others to create a consortium with sufficient demand to attract competitors and thereby negotiate lower rates or at least secure efficiencies, particularly in lower density regions."  For 20 years, the FCC has been saying "bigger is better," and now Commissioner O'Rielly is asking an applicant to explain why bigger is better.

Monday, July 15, 2019

Mystery extension

Huh?  I was just reading the cover letter from one of the Commission's "Streamlined Resolution of Requests Related to Actions by [USAC]" and I noticed this sentence: "If the Bureau has dismissed or denied your appeal and you would like to seek reconsideration of that decision, the deadline to file a petition for reconsideration or application for review by the full Commission is 116 days from the release date."  What?! 116 days?  We were told it was 30 days. The FCC has denied a Request for Reconsideration that was filed 32 days after the decision.

Did the rules change?  What are the rules?  Well the above-mentioned cover letter provides the relevant chapter.  Here we go, 47 CFR § 1.106 (f): "The petition for reconsideration and any supplement thereto shall be filed within 30 days...."  So there, 30 days. 

Can anyone tell me where the 116 days comes from?  That's a really odd number.

Not that I'm complaining.

Wednesday, July 10, 2019

Let's hear from an overbuilder

In a new twist on the overbuilding debate, an Alaskan telecom provider has filed with the FCC asking for changes to E-Rate and Rural Healthcare rules, because they're finding it difficult to compete under the current rules.  The telecom costs in Alaska are thrilling, so I thought I'd see what's up.

This request is kind of the flipside of the Texas Carriers case: this time, the overbuilder is complaining that the owner of existing fiber has an unfair advantage.

Before I get into the particulars, here's an overbuilding success story from an article written the last time Quintillion tried to get the FCC to put the squeeze on GCI:
Crawford cited the Nome School District as an example. The district has five schools, and 700 students, and paid $305,000 per month to GCI for its internet service.
According to Crawford’s letter, the district was able to reduce its bill to $95,000 a month once Quintillion connected to the shore from its subsea fiber optic cable.
A little overbuilding, and the price drops by 69%.  If Commissioner O'Rielly's rules had been enacted before Quintillion laid cable in 2016, then Quintillion would have had to use GCI's infrastructure, and the Nome schools would be paying 3 times as much.  At a cost to the E-Rate of $3 million per year.

Who are Quintillion and GCI?

Quintillion, the complainant: a company that has laid oceanic fiber around the northwest corner of Alaska as part of a plan to build a link from London to Tokyo.  )Seems crazy until you see the map:
OK, it still seems crazy.  I believe the idea is to shave 24 ms off the current 170 ms latency between London and Tokyo; high-speed traders will pay a lot to give their trades that speed boost.)
Here's a map that just shows Quintillion's Alaska network:

GCI: a company that has created a fiber-microwave WAN that covers much of Alaska:

So Quintillion complains:
  1. GCI gets more than 75% of USF funding awarded in Alaska.
  2. The FCC cut GCI's Rural Healthcare (RHC) funding by 26% in 2017.
  3. Roughly 50% of funded E-Rate Internet commitments in Alaska received only one bid.
  4. GCI’s market dominance resulted, in part, from ... $44 million federal BIP [Broadband Innovations Program] grants.
  5. GCI insists on using its own network, refusing to use Quintillion's.
To which I say:
  1. That is troubling.  Unless, like Commissioner O'Rielly, you believe that since federal funding helped pay for GCI's network, it should be protected from competition.  Me?  I say: "Overbuild, Quintillion, overbuild!"  The consensus up in Alaska is that a monopoly is not a good thing.
  2. I'm too lazy to look into that, but I'll bet that has something to do with RHC rules about the cost of services in rural areas compared to costs in urban areas.  And, of course, it's a result of GCI having infrastructure where no one else does.
  3. That is troubling.  Some might say it is a good sign, that there is no wasteful overbuilding, but me, I think it shows a disturbing lack of competition.
  4. And therefore, according to Commissioner O'Rielly's reasoning, the government's $44 million investment should be protected from any competitor building service to any locations covered by their network.
  5. GCI's cost to use their own existing network is close to $0.  Quintillion, did you offer to let them use your network for close to $0?  Keep lowering your price, and eventually they'll stop refusing.
And what solutions does Quintillion propose?
  1. Make changes to the RHC:
    1. Consider more cost-effective middle-mile and backhaul solutions.
    2. Extend the bid period to 90 days.
    3. If only one bid is received, limit the contract length to one year.
    4. Allow more flexibility on changing service providers for single-bid awards.
    5. Audit single-bid awards.
  2. Make changes to the E-Rate:
    1. Extend the bid period to 90 days from 28 days.
    2. Require single-bid awards to submit cost and rate information that will be made public.
    3. If only one bid is received, limit the contract length to one year.
    4. Audit single-bid awards.  
To which I say:
  1. Not my circus, not my monkeys. 
  2. OK, this is my circus ("my circus" meaning "I belong to this circus," not "this circus belongs to me"):
    1. No.  What will service providers be doing for 90 days?  How many more bids will come in?  I agree with letting service providers have more time in some cases, but not 90 days in all cases; let local officials determine the appropriate amount of time.
    2. That info is on the Form 471, which is public.  And GCI has already made their rates public.  And we're talking about public bids here: file a FOIA request and get the whole bid if you want.  That's 3 ways you already have to get the information for all bids, not just single-bids.
    3. Really?!  You're going to make schools and libraries go through the formality of re-applying and getting a single bid year after year?  In most Internet contracts I've seen, the price drops sharply if you sign up for at least 3 years.  So schools and libraries have to pay more and go through more hassle, just because one of these years, some other company might want to bother bidding?
    4. How about this?  In single-bid situations, we audit all the telecom companies in the state who didn't bid?  Why go after the applicant and the service provider who participated in the competitive bidding process?  Instead, let's investigate service providers who didn't, especially any that have gotten any federal funding.
Why did Quintillion make this filing?  Quintillion is a wholesaler.  They sell middle-mile to service providers.  They brought a fiber connection to five towns in Alaska, but sell only to other telcos, who then sell to consumers (including school districts).  Why are they complaining?  Have they ever participated in a bid?

There are only 3 school districts that Quintillion fiber would serve: Nome, Northwest Arctic Borough (Kotzebue) and North Slope (Port Hope to Prudhoe Bay).  Quintillion already won Nome, but let's see:  Northwest Arctic Borough is in a $6-million-per-year contract with GCI until 2021 (with voluntary extensions through 2023); North Slope is in a $7-million-per-year contract until 2021 (with voluntary extensions through 2031).  OK, I can see why Quintillion wants to give their resellers an opportunity to figure out how they're going to connect the schools to Quintillion's POP in those towns.  But hey, if the resellers start planning now, they won't need 90 days after the Form 470 goes up.  Then in the fall of 2020, let the districts know how much money you can save them, and they'll bail on the contract extensions and you'll win the business for FY 2021-2022.

So what do you think?  Is Quintillion an evil overbuilder, hurting the value of the federal government's investment in GCI's network?  Or are they a competitor prevented from bringing costs down for schools? (And, since the E-Rate is paying 80-90% of the cost, bringing down costs for the federal government.)