Search This Blog

Thursday, May 28, 2015

Some pig!

Check it out: the FCC is redoing their website.  Here is a prototype.  I like it, though I don't often hit the homepage.

Have a suggestion to improve it?  Send it to

Maybe I should start a spamment campaign to bring back hraunfoss.

Eligibility brief

The new proposed Eligible Services List came out last week, but I'm just getting around to reading it today.  I'll give you my reactions as I read though.  (It's sort of like one of those unboxing videos, only more boring.)

Before we open the box, let's just look at the size of it.  4 pages of pre-amble, 8 pages of ESL.  And shaking the box a bit, I see that the ESL rules are rattling around in mostly empty pages. Uh oh, looks like we're stuck with ESL Lite for another year.

The card attached to the box is not bad: the proposed ESL was released on May 21st, with comments due by June 22nd.  A month for comments gives us all a long time to ruminate.  And the early due date for comments means we might get an ESL in the summer.  That's only fair, since at the recent fiber WAN workshop, the FCC wanted applicants to get 470s up in July.

The package has a meager ribbon: In the past, the ESL was always in Docket 02-6.  Somewhere along the way, Docket 09-51 got thrown in.  Last year, they added Docket 13-184, the docket created for the E-Rate modernization NPRM.  This year, we're down to Docket 13-184.  Is Docket 02-6 going the way of the dodo?  Why not a new docket for the proposed ESL each year?  We need to do something to address the E-Rate docket deficiency.  The single docket is good news for paper filers, since it means they only need to send one copy along with the original, not the 5 copies they had to send when 3 dockets were listed.

Oh, dear, I don't like the pattern on the wrapping paper: the small "r" in "E-Rate."  At least the use was consistent.  And again the phrase "more commonly referred to as the E-rate program," when as I've said before, the capital "R" is more common among Congresspeople, the press and FCC Commissioners, and even the Chairman (at least until recently).

First smile comes from a footnote: "...we do not invite requests for reconsideration of the E-rate Modernization Orders as part of this notice seeking comment on the proposed funding year 2016 ESL."  In other words: "Don't tell us you want Web hosting and voice back.  Tell us what you think about the wording and format."

"The proposed ESL adds equipment necessary to make a broadband service functional to the list of eligible costs for leased lit fiber, dark fiber, and self-provisioned broadband networks. This clarification is necessary to fully equalize the treatment of lit and dark fiber and to support self-provisioned broadband networks."  Yes, and combined with the earlier "...applicants must seek competitive bids for network maintenance and operation, and all other eligible services and equipment," it means that dark-fiber companies can't make their bid look cheaper by failing to include the cost of electronics.  A win for the phone companies.

"...the costs for bundled voice and data services provided over a single circuit, must be cost allocated."  That's got me worried.  Am I going to have to make up a number for the miniscule portion of my 100 Mbps circuit that is being used for VoIP, and then apply the lower voice discount to that?  We'll see when I get to the actual ESL.

"...applicants that may be receiving ISDN as bundled voice and data service...."  Nope.  ISDN is a data service.  It's just that no one uses it for anything but voice.  (Is there anyone out there still using 3 BRIs for videoconferencing?  Please tell me no one is using a BRI to dial up the Internet.  And a PRI is just a T-1 set up for voice, so if someone's using it for data, it's just a T-1.)  Also, at least around here, unless the PRI has been there for more than 3 years, when a provider gives an applicant a PRI, it only stays a PRI until it hits the service provider's router in the applicant's building.  It leaves the back end of that router as IP packets.

Hey, look, comments to be filed at!  Faithful readers will know that I miss those icelandic waterfalls.  Alas, it just forwards you to the real page:  Stupid Web trick: For those of you that miss hraunfoss as much as I do, you can use  Svartifoss does not work, so I guess the FCC never took my suggestion to use that name.

"Additional guidance from USAC about the E-rate application process and about eligible services,
including a glossary of terms, is available at USAC’s website at  The documents on USAC’s website are not incorporated by reference into the ESL and do not bind the Commission.  Thus, they will not be used to determine whether a service or product is eligible."  So you want to know what is meant by "Radio loop" or "Interconnected voice over Internet protocol"?  Go look on USAC's Web site.  But that definition is not binding, so you're really just guessing.

Hey, where is that glossary?  I looked in the Reference Area: nothing.  I searched for the word "glossary" and the only two uses of the word were here and here, with both pages saying the glossary was part of the ESL.  So the ESL says the glossary is on the USAC website, and the USAC website says it's in the ESL.  Beautiful.

OK, let's see which digital transmission services get us to 1) broadband (defined by the FCC as 25 Mbps) and 2) 100 Mbps (the goal set in the E-Rate Modernization Order).
Technology Top Speed Broadband 100 Mbps
Asynchronous Transfer Mode (ATM) 10 Gbps Yes Yes
Broadband over Power Lines (BPL) Service not
? No
Cable Modem 150 Mbps Yes Yes
Digital Subscriber Line (DSL), theoretical
Digital Subscriber Line (DSL), actual
100 Mbps
6-50 Mbps
DS-1 (T-1)
DS-3 (T-3)
Fractional T-1
Fractional T-3
1.5 Mbps
45 Mbps
< 1.5 Mbps
< 45 mbps
Ethernet 10 Gbps Yes Yes
1.5 Mbps
128 kbps
Leased Lit Fiber 10 Gbps Yes Yes
Dark Fiber 10 Gbps Yes Yes
Self-Provisioned Broadband Networks 10 Gbps Yes Yes
Frame Relay 45 Mbps Yes No
Multi-Protocol Label Switching (MPLS) 10 Gbps Yes Yes
50 Mbps
150 Mbps
622 Mbps
n x 50 Mbps
if n>1
Satellite Service, theoretical
Satellite Service, actual
1 Gbps
15 Mbps
Switched Multimegabit Data Service 45 Mbps
(if you can travel back to the 1990s)
Yes No
Telephone dial-up 0.056 Mbps No No
Wireless services (e.g., microwave) 1 Gbps Yes Yes
So if we tossed voice out of the program, is it time to start tossing digital transmission that doesn't help us get to the goal of the program?  Or at least get rid of circuits that aren't even broadband?

"Applicants may seek special construction funding for the upfront, non-recurring costs of deployment of new or upgraded facilities, including design and engineering, project management, and construction of network facilities."  Does that mean I can get E-Rate funding for the cost of a design firm to write my RFP?  Or is it more like the design of a C2 solution, which has to be done by the installer right before installation?  If the FCC wants top-notch fiber RFPs, they should pay part of the cost to create them.

"The reduced discount rate for voice services will apply to all applicants and all costs for the provision of telephone services and circuit capacity dedicated to providing voice services...."  Hmm....  "Dedicated," eh?  Well, you can't really "dedicate" a portion of a circuit in a packet-switched network, so I guess this only applies to dedicated circuits.  That's OK.

"Firewall protection that is provided by a vendor other than the Internet access provider or priced out separately will be considered a Category Two internal connections component." By the people who brought you the On-Premise Priority One Equipment morass, it's the brand-new "Off-Premise Category Two Service," coming soon to an ESL near you.  I guess we already had O-PCTS with cloud-based WLAN controllers.  I dislike loopholes, but I have to say that O-PCTS is a better idea than O-PPOE, because no one's trying to cram things into C2.

"Access points...such as wireless access points."  That implies that there are access points that aren't wireless.  If such a device were not wireless, it would be called a "hub."

Hey, UPS is just listed as eligible.  There is absolutely nothing in the ESL which says a UPS is only eligible if supports eligible equipment.  So go ahead and plug your ineligible servers into that E-Rated UPS.

"A manufacturer’s multi-year warranty for a period up to three years that is provided as an integral part of an eligible component, without a separately identifiable cost, may be included in the cost of the component."  That means that if the warranty is longer than 3 years, it may not be included in the cost of the component.  So now I've got to make up some cost for the lifetime hardware warranty that's standard on most manufacturers' switches?  Oh, wait, if I didn't put the warranty on my 470 or RFP, then I can just say it's ancillary.  What if I ask for at least a 3-year warranty and I end up with a lifetime warranty?  I think I'll need a lawyer to split the hair of whether asking for an eligible warranty but getting an ineligible one still fits into the Ancillary Use definition.  Or since it's a bundled warranty, is it eligible, just not as part of the purchase?  Three years after buying the equipment, I could start making separate BMIC FRNs for $0 to cover the non-cost of the bundled warranty....

Why do I think some service provider is going to try to jimmy their video server into the Caching definition?

"The agreement or contract must specifically identify the eligible internal connections covered, including product name, model number, and location."  Why?

"Upfront estimates that cover the full cost of every piece of eligible equipment."  That implies that as long as my upfront estimate is for less than the full cost of every piece, I'm good.

And that's all she wrote.

Wait!  We just heard last week that the FCC will be requiring RFPs for self-provisioned fiber, and requiring that applicants seeking self-provisioned fiber have to solicit bids for leased fiber.  How is that not mentioned?  That's a whopping new and unusual condition on eligibility.  It should be in the ESL.

Does anyone else feel like this blog post is longer than the ESL?

Wednesday, May 27, 2015

Success? Well, yes, but...

Mostly, I agree with Chairman Wheeler's latest blog post.  But of course, I'm going to focus on areas of disagreement.  It's sort the inverse of the new ESL: I'm not going to list everything I agree with, so if something he said isn't mentioned in this blog post, you can assume I agree with it.  Unless I have disagreed with it somewhere else in this blog or filings with the FCC.

First, the most critical issue: the Chairman used a small "r" in spelling "E-rate."  No, no, no.  As recently as September, he was using the big "R."  What happened?

Next, a disagreement that seems nitpicky, but is kind of important.  The Chairman talks of "supporting cutting-edge, one-to-one digital learning."  From what I've seen, schools have blown right by 1:1 and are passing 2:1 and wondering if it will stop at 3:1 (phone, laptop/netbook and tablet/gaming device).  It's important because it affects the recent excitement about "duplicative" services.  Because a lot of schools aren't going to want student smartphones using the same Internet connection as online testing.  So we'll increasingly see schools buying a separate Internet connection for the students' second (and third) devices (and the Internet connection for testing is actually going to be 2 connections, because if testing gets paused by connectivity problems, it's in the papers).  Unfortunately, the current duplicative policy is based on the mistaken belief that each student has at most one device (and the mistaken belief that a school can rely on a single connection).

I don't see the "pent-up demand."  Here's a table showing what I mean:
Disc. Max. Demand Requested % Requested
20% $ 6,049,500 $ 531,458 8.79%
40% $ 338,603,400 $ 39,027,526 11.53%
50% $ 444,421,125 $ 68,381,954 15.39%
60% $ 628,483,860 $ 141,881,622 22.58%
70% $ 1,091,851,950 $ 64,178,921 5.88%
80% $ 1,833,255,480 $ 1,351,202,599 49.71%
85% $ 885,083,963
Total $ 5,227,749,278 $ 1,665,204,080 31.85%
The "Max. Demand" column shows the maximum demand for each discount band, from a spreadsheet I made back in June.  The "Requested" shows what applicants actually requested, based on USAC's Demand Estimate.  It's a bit of a blunt instrument, but it does show roughly how voracious demand was.  Except for high-discount applicants, I don't see a release of pent-up demand.  The 80-90% applicants went after it a bit, but requests have been over $2 billion in recent years, so I don't think $1.3 billion is big.

Ooh, some example applicants, those are always fun.

How did Kindred, ND get selected?  I just wonder what made them the poster child for small rural C2 requests.

The application hasn't been through PIA, so I'm not going to discuss what I think of the equipment in the request or what it costs, but I do want to point out one thing.  According to the 471, the Kindred school district has 689 students, giving them a C2 budget of $103,350.00, so the total pre-discount charges of $98,725.00 fits in nicely right?  Nope.  Unfortunately, the high school only has 299 kids, so it only has a budget of  $44,850.00, and since the district opted to split the cost evenly between locations, each location has a pre-discount cost of $49,362.50.  So right away, PIA is going to knock $4,512.50 off the pre-discount amount, which means Kindred loses $1,805.00 in funding because the FCC decided that $150-in-5 budgets have to be divided by location.  In general, does anyone think that an elementary school with 390 kids really needs 30% more C2 funding than a high school with 299 kids?  Because that's what the rules say.

And you know what Kindred's E-Rate funding will be in 2 years?  $0.  Because their C2 budget is almost used up, and their voice discount will drop to 0% by then.  OK, so they'll still get funding indirectly as part of the big ND School Net consortium, but they will no longer be filing a 471.  Actually, I guess that's what the FCC wants: everything through a consortium.

The Philadelphia School District gets mentioned, too.  Yes, that $5.7 million will help connect their kids to Wi-Fi.  But in the 5 funding years before that, the average annual request was $6.3 million.  And the overall average since the start of the program is $16 million per year.  So $5.7 million is only exciting if you're thinking of the last 2 years, when the FCC decided not to fund any P2.

Philly shows us another dark side of the per-location $150-in-5 budgets.  Open their application, go to the Item 21 information, and expand everything.  The relatively simple purchase of some wireless gear has exploded into a bookkeeping nightmare.  Imagine an auditor trying to figure out if the amount listed for each school was used in that school.  And what did we gain from all that complexity?

I have to admit, though, that the FCC's demand projections were better than my doom-and-gloom.  And the reforms are driving funding to broadband and Wi-Fi.  So overall, I think the Chairman has a right to crow.

Monday, May 18, 2015

Bundling a loophole

I thought Managed Internal Broadband Services (MIBS) was a messy loophole through which service providers will try to force all sorts of things.  Here comes a prime candidate: Sprint has launched a service that seems to be called OfficeFuel or WorkPlace-as-a-Service (WPaaS).  It is a full-service MIBS offering, but since it's Sprint, of course it also includes voice and mobile data.  As if figuring out MIBS weren't difficult enough, now we're going to get differently eligible services.  The cost allocation possibilities are extreme. The voice portion is eligible, but at a reduced discount.  And hey, if Sprint is running the whole show, they'll be able to show what portion of your cell phone's data goes through Wi-Fi, and since the cost of your cellphone data plan will be part of the monthly per-user fee, will mobile data creep back in?

Ugh, MIBS.

Monday, May 11, 2015

You celebrate. I'll grumble.

OK, so maybe I was a little pessimistic.  My pre-window estimate was that Category Two (C2) funding would run out somewhere between 70% and 50% applicants.  Well, last week USAC released the official Demand Estimate, and the total demand is just over the $3.9 billion cap.  Then 2 days later, the FCC issues an order directing USAC to cover it all.

Now it's official: everybody who applied correctly will get C2 funding in 2015-2016.  That is very good news for the stability of the fund.

If demand was higher than the cap, how did they fund everyone?  The rollover, of course.  Because the FCC has chosen to violate their own rule and deny all C2 funding for the past 2 years, USAC is sitting on $1.6 billion in unused funds.  So they can cover the $0.019 billion over the cap.

But the way they've chosen to cover it is bad for the stability of the fund.  To maximize the safety of the fund, the FCC should have raised the contribution factor to pull in the cap, $3.9 billion, and then used just a smidgen from the $1.6 billion unused funds war chest.  Because it is possible that demand next year will be way over $3.9 billion, and the FCC will be wishing it still had that unused slush fund.

OK, that seems extreme, but I think they should have left some of the money in the unused pile.  For example, they could have said, "OK, here's that extra $1 billion that the Chairman promised for FY 2015," and left $600 million in the unused pile for next year.

What did the FCC do?  They said: "We ... direct USAC to fully fund eligible category one services under the new cap. We also direct USAC to fully fund eligible category two services, first using as much as $1.575 billion in E-rate funds unused from previous years, and then using any additional funds needed under the new cap to fully meet demand."

Seems reasonable, right?  No, it's actually a mistake.  Maybe a big one.  Walk with me as we go through what will happen.

First, USAC will fund Category One (C1).  Total C1 requests are $2.25 billion.  USAC usually ends up denying/reducing a little over 10% of the total requested amount.  So USAC will probably approve something like $2 billion in C1 funding.

Next, USAC will start using the $1.6 billion in unused rollover funds to cover P2.  Total C2 requests are almost $1.7 billion.  Applying a 10% denial factor, we get to about $1.5 billion in approvals.  Hey, we didn't eat up the whole rollover!

Did you notice the problem?  The amount that E-Rate pulled from the Universal Service Fund (USF) for FY 2014-2015 was just over $2.4 billion, the program cap.  But in 2015-2016, the E-Rate is only going to pull enough to cover C1, which is going to be something like $2 billion.  Now normally, I would say it's a good thing that the E-Rate program will be reducing the contribution factor.  But does anyone believe that demand is going to decrease next year?  Does anyone believe that the FCC is going to find another $1.6 billion for FY 2016-2017?  Remember that the big reserve is the result of no C2 funding for 2 straight years, and by leveraging the fund.  Neither of those options is available in the future.

So next year, it is possible that the E-Rate will need to pull $3.9 billion from the USF. Since we're only taking $2 billion this year, that almost doubles the E-Rate's demand on the USF, which means the contribution factor is going to spike.  This year's "yippee! we cut spending" party will give us an "oh my aching contribution factor" hangover next year.

And the spike will hit in July 2016, just a few months before the elections.  If I were Hillary Clinton (or any Democratic candidate for Congress), I would be on the phone to Chairman Wheeler right now.  And I would be saying, "Take $3 billion or so from the fund this year, and let's have the 'yippee! we cut spending' party at this time next year."

Meanwhile, Commissioners Pai and O'Rielly should start resting their voices now, so they'll be ready to give a proper outcry at this time next year.