[Starlink] The DoD "Transport Layer"

[David P. Reed]

Ah, hell. DoD ain't the ARPA of the '70's. The Beltway Bandits aren't gonna build any friggin' fairy Internet. Nope, Space calls for John Wayne-style Gold Plated designs. No COTS architecture is relevent to "space, the final frontier".
 
This is gonna be pure "5G" (the protocol stack of that name, requiring algorithms of that name).
Because, after all, Space Force wants to argue to control a bigger budget and develop its own unique "Space Transport" system.
 
Now that is sarcastic, but since I was involved in the rationale for the Internet *in the ARPA concept*, and know Vint and Bob's mandate, I think it will turn out to be more true than not.
 
The reasons the branches of DoD cannot interoperate their networks is basically caused by "contracting out" based on bids with specifications that allow each vendor to throw non-interoperable features into the mix, so that they lock in DoD to lots of incompatible technology. The bureaucrats (Admirals, Generals, Colonels, and civilian agency staff) only get to "review proposals" for "complete systems".
 
The most recent example of such a fiasco I happened to get close to was the attempt to modernize the radio "waveforms" (really protocols) of DoD spectrum. OMG. What a procurement nightmare. The clear career path of those in charge of procurement was to "retire" and go to work for the winning contractors as employee or consultant to the sales process.
 
A less recent example was when services came in Iraq when it was learned that systems never interoperated - why? Well, some young officers in their "spare time" hacked together a system called RIPRNET, which was basically the Internet architecture using cheap COTS walkie-talkies as links that bypassed all the gold plated but isolated tech. (and of course the robots deployed in Iraq spoke IP, as well).
 
Is this "corrupt"? No, it's just profit maximizing for the vendors. Free market if you don't plan for what will be needed.
 
There certainly won't be any "open source licensing" on any software here. The trivial argument that the Chinese will infiltrate the design is a pocket veto.
 
The real question is why doesn't this  network get built using IPv6 transport? It would certainly be a lot cheaper and it would work from day one. Ask your Congressional representatives, why? But having worked on the FCC TAC and so forth, as well as dealing with some smart folks in DoD over the years (like Vint and Bob, but also Admiral McMullen, and John McCain's staff) the technolobbyists have already circulated bogus technical reasons why the Internet cannot work in Space. Is there an Internet in Star Trek? Well, that proves it then! There are lots of consultants who worked for proprietary gear vendors who will write impressive but obscurantist technical papers to argue that the Internet architecture can't work at all in Space. And Congressional staffers who are "in on the con" will wave those papers around.
 
How did we get the Internet? NOT because DoD spec'ed it into contracts for DoD procurement. Nope, a small group invented internetworking as a universal *overlay* at the lowest levels of hardware, and paid BBN and some universities to demonstrate how to make it work. That was cheap - not the kind of money that a "real" Space Force must spend!
 
 
> Date: Fri, 14 Oct 2022 06:59:46 -0700

> From: Dave Taht <dave.taht at gmail.com>
> To: Dave Taht via Starlink <starlink at lists.bufferbloat.net>
> Subject: [Starlink] The DoD "Transport Layer"
> Message-ID:
> <CAA93jw6RReXwaa0DvfkXNnHjfzxv+3th2p7qygEwd98srLgiZw at mail.gmail.com>
> Content-Type: text/plain; charset="UTF-8"
> 
> It looks like an "internet" in space is shaping up. In a much earlier
> thread on this list, there was a presentation about the moon<->earth
> architectures and standards being sorted out. It was an awful looking
> amalgam of every technology we have available today from lte to wifi
> to the bundle protocol and everything in between. I can't find the
> site that pointed at the specs now (?)
> 
> Right now it's reminding me of the bad ole days, where you'd find me on
> 
> ..leo!LunarGW!rhysling_crater!dtaht.
> 
> And then there's this that went by yesterday:
> 
> "DIU’s hybrid space architecture would use commercial communication
> systems as transport pipes to move data collected by imaging
> satellites and deliver it quickly to government users. The concept
> assumes that commercial satellites will talk to each other via
> interoperable links.
> 
> Shimmin said his office awards contracts to commercial companies with
> incentives that “gently encourage different vendors to cooperate
> together.” By doing that, “we’ve created a much more
> collaborative
> relationship with our vendors.”
 
[DPR] contract requirements create collusion among winners, not interoperability or inexpensive upgradeability. There's a better strategy. Not that I can be heard on this, I suspect even Vint can't be heard on this:
Mandate that Internet Architectures MUST be used unless a panel of independent scientists who have no financial or careerist conflicts can be convinced it can't work.
Mandate the End-to-end argument be used to place functionality at the "edges" where updates can be incrementally made without changing the underlying transport at all.

> 
> DIU is working with the Space Force and the Air Force Research
> Laboratory on the hybrid architecture. The project is intended to
> support Pentagon efforts to connect ground, air, maritime and space
> systems, a concept known as Joint All-Domain Command and Control, or
> JADC2.
> 
> The backbone of the hybrid network will be DoD’s Transport Layer, a
> constellation to be deployed by the Space Development Agency, he said.
 
[DPR] Oh great, a proposal to scrap all of the past and replace it with a grand unified theory of the future that will have to completed 100% before it can be tested. This may seem like the Internet, but in fact, the Internet was the opposite - a modest proposal that could be tested from day 0.

> “We want to augment it with the commercial communications
> architectures that are coming online to proliferate the internet in
> space, get every satellite talking to every other satellite, relaying
> through ground stations regardless of who owns the ground stations,
> they should all function as routers.”
> 
> The thinking in JADC2 is “to embrace multiple providers so we don’t
> have a single point of failure,” said Shimmin. "
 
[DPR] Conflating multiple providers with fault tolerance is a terrible mistake. It's not the use of multiple providers that creates the fault tolerance of the Internet Architecture. It's that it *doesn't matter if a provider doesn't deliver a packet*. That's not a procurement issue, but a routing issue. That is, you don't start by choosing providers. You start by ignoring providers proposals. Instead you create an overlay that is provider independent, specify it, and then accept bids to extend the overlay more broadly.
 
> 
> 
> https://spacenews.com/starlinks-market-dominance-affecting-dods-hybrid-network-plans/
>
[DPR] C'mon. Starlink doesn't have "market dominance" in any market. It may have more satellites in space, but that's because satellites are a dime a dozen. DoD clearly hasn't looked at the space communications issues. SpaceX may have launch capability market dominance.
 
> 
> 
> --
> This song goes out to all the folk that thought Stadia would work:
> https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-6981366665607352320-FXtz
> Dave Täht CEO, TekLibre, LLC

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/starlink/attachments/20221014/1af760d2/attachment.html>