From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-yb1-xb2f.google.com (mail-yb1-xb2f.google.com [IPv6:2607:f8b0:4864:20::b2f]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 9E8A23CB40; Wed, 29 Mar 2023 15:37:13 -0400 (EDT) Received: by mail-yb1-xb2f.google.com with SMTP id e65so20749753ybh.10; Wed, 29 Mar 2023 12:37:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; t=1680118633; h=cc:to:subject:message-id:date:from:in-reply-to:references :mime-version:from:to:cc:subject:date:message-id:reply-to; bh=Qc7R1LCC06hyg+5foNQAJ1B0EGH+MH038Db1QEZ7m4A=; b=OAVGpL7gn2o+JUQwQjkbJjmuF1ySzWugzLx3abxTxq9dCOsNDiARcNmzqAchCNruVC We7FtimelDDlv1r34PatI2JvQryqohHUmLRJHe3GgLoF1yjW/4EexWCMdr9q+D3TTIBr QJi/ISv+pJMd1OYJEsDbRfznc1l63rAfu+4fY69Vso2RcHX4hqLu3hvidY/I1spB5mFW Hc1oXN+ghPFfFGS4S1/IN7K4+BClvp9oxUNhYeknLEQpdufgfPADIuJ00Sz8R3w4Sshb ZvpfSgi451hlzrhhtjSeCqj9jajxGXX0gnH3FHnokn73/VVaR56DJOvM0qu37Qakw60f 5xHA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; t=1680118633; h=cc:to:subject:message-id:date:from:in-reply-to:references :mime-version:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=Qc7R1LCC06hyg+5foNQAJ1B0EGH+MH038Db1QEZ7m4A=; b=zniiSkjjKsjJka/joGFt6ibqC74xpTIqc8Xd/b/R3M4TehHZzJ1lWAXW78aFqBB8+2 OSq3uFxil4Mo6hho0YangbdLF0RTnxTkZ17AffLU8XJGhOlIUyB9aakOMbq9pgqgQPKI ph/JBX19ZT9KHKefUpt73DX+3q4MFb46VBAHtsySFpxp8cbCKe2Yf3Tprr1peEgEl7zs C6w84U4eEqvFqceGy7A5Ucjm+zDKFDwtsLJyx15av7iLgbVzto+AD9Ojh/ptfH4i6qfI y1EKLDdLMrctlvBftrN+/A6UUQ8PQsh6RrV52JnVAT4w1H9McFvriqaGJOBcwhVyFZkc W6EA== X-Gm-Message-State: AAQBX9dVM4BklZZ+QFsII6SG3tZ1hjahDmIGtjGqeMkJjl4y1CMtQyB0 BB5cgSCBYGRC2WYL/+bhisxhCTQ4sJ2q1I7aSUw= X-Google-Smtp-Source: AKy350Y/aqmrvBXm76RsmERGVnO1d5VVIp//gyvDt26Sgn12TP3WBfDEN58Ttu1lMVgPVidYlOaySh0z0zCRarlosEM= X-Received: by 2002:a25:bb85:0:b0:b7c:1144:a708 with SMTP id y5-20020a25bb85000000b00b7c1144a708mr7027834ybg.12.1680118628129; Wed, 29 Mar 2023 12:37:08 -0700 (PDT) Received: from 1064022179695 named unknown by gmailapi.google.com with HTTPREST; Wed, 29 Mar 2023 15:37:07 -0400 Received: from 1064022179695 named unknown by gmailapi.google.com with HTTPREST; Wed, 29 Mar 2023 15:37:04 -0400 Mime-Version: 1.0 (Mimestream 0.41.6) References: <1d6c10c9a692bb3f2869fb1b40fa449a@rjmcmahon.com> <569691b3e7dfc57bbf98c4fc168fc6cf@rjmcmahon.com> <2885829.1679221616@dyas> <20230321001019.GA4531@sunf68.rd.bbc.co.uk> <4295238B-FA57-49B6-B57B-78FFB2603B90@gmx.de> <8301258b8fffa18bd14279bff043dd03@rjmcmahon.com> <43bcbc338aecb44a1bef49489ab6f9c8@rjmcmahon.com> <60e70b637df76234639780ab08f25d82@rjmcmahon.com> <9edd011a1a6615470b34e0837896a15f@rjmcmahon.com> <6EB62755-EF23-44BA-B2FF-66FAC708653D@gmx.de> <6qnq34os-3qss-s4q7-s286-2s49q890q920@ynat.uz> <27aea5070eeb1b1535f3e75489295feb@rjmcmahon.com> <08526EAC-7EA3-4BFA-A231-B2935E09C8AC@gmx.de> In-Reply-To: From: dan Date: Wed, 29 Mar 2023 15:37:07 -0400 Message-ID: To: rjmcmahon Cc: Larry Press , David Lang , Frantisek Borsik , libreqos , Dave Taht via Starlink , bloat , Sebastian Moeller Content-Type: multipart/alternative; boundary="00000000000002ae6805f80f1a65" Subject: Re: [Starlink] [Bloat] On fiber as critical infrastructure w/Comcast chat X-BeenThere: starlink@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: "Starlink has bufferbloat. Bad." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 29 Mar 2023 19:37:13 -0000 --00000000000002ae6805f80f1a65 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On Mar 29, 2023 at 1:02:51 PM, rjmcmahon wrote: > Hi Sebastian, > > I'm fine with municipal broadband projects. I do think they'll need to > leverage the economy of scale driven by others. An ASIC tape out, just > for the design, is ~$80M and a minimum of 18 mos of high-skill, > engineering work by many specialties, signal integrity, etc. Then, after > all that, one has to get in line with a foundry that needs to produce in > volume per their mfg economies of scale. These markets fundamentally > have to be driven by large orders from providers with millions of > subscribers. That's just the market & engineering reality of things. > > Every ASIC necessary to deploy is already on the market in high volume. No additional ~$80M needs spent. ~$80M that MUST come from the customer at the end of the day. Another increase in broadband costs. Every massive change you suggest will pull money from actually running mainline fiber to communities where various technologies can already deliver huge speeds at low latency. I=E2=80=99m an operator, my primary limitations logistically = speaking is inability to get 10Gbps+ fiber off the existing fiber footprint. Even the lowly DSL footprint could be upgraded with relative ease to get a few hundred Mbps if, and forgive me for leaning not his so hard, the previously designed monopoly that owned not only the copper plant but also the fiber that is already there wasn=E2=80=99t waiting around for the next government= hand out before upgrading. Fiber to the DSLAM and VSDL would be a nearly instant upgrade to 100+ x 50+ speeds for easily 80% of rural users. We don=E2=80=99t need a completely different model (FiWi) when we have all = of the parts and pieces in mass production and available right now, we have a political system that promotes monopoly and actively encourages them to wait until either a self funded competitor moves in or government money shows up with mandates. There is no reason at all to have 3-7Mbps DSL in most of America. This is not a technical limit. An aspect of the FiWi argument is that these NRE spends today and > tomorrow are mostly from SERDES & lasers/optics in the data centers and > the CMOS radios & PHYs in handsets. Let us look here for the thousands > of engineers needed and for the supply of parts for the next decade+. I > don't see it coming from anywhere else. > We have 100G hardware routers from multiple vendors, Qualcomm, Broadcom, Marvell. We have 1-100G optics on the market today for cheap. Marvell makes a line of chips that can do 40Gbps hardware switch or routed for like $20, get=E2=80=99s put in $200 MikroTik devices today. A grand gets you int= o a device that can do 100G today. Obviously that=E2=80=99s from the cheapest = vendor but 2-10x that price will get you into the =E2=80=98good stuff=E2=80=99. W= e already have this. > Then we need the in-premise fiber installers and the OSP labor forces > who are critical to our success. > > And finally, it's the operations & management and the reduction of those > expenses in a manner that scales. > > Where exactly are the costs, operations, and management savings here? Basically this leads me to the question which I=E2=80=99m asking with an at= tempt to avoid condescension, do you/have you run an ISP? My operations and management costs are primarily customer service and logistic (vehicles, labor, and so on) and not network management. Fiber in-premise has a negative value. It=E2=80=99s more expensive to term= inate and repair, port costs are more, vastly (like 100x) more likely to damage a fiber patch cable vs cat5e, and the advantages of fiber are lost on short distances. 1,2.5, 5, and 10G copper is easy, cheap to terminate, cheap to install, cheap ports in switches, cheap ports on devices, and fast. The entire =E2=80=98need=E2=80=99 for fiber in this context is the FiWi concept= of centralized networking which again IMO is something ALL IT/MSP will outright reject killing it off for business uses and will not fare well for consumers who are concerned more and more about privacy. Just my opinion here, but the entirety of the FiWi concept will be dead on arrival with almost all opposing it and only a few supporters. --00000000000002ae6805f80f1a65 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable





On Mar 29, 2023 at 1:02:51 PM, rj= mcmahon <rjmcmahon@rjmcmahon.= com> wrote:
=20
Hi Sebastian,

I'm fine with municipal broadband projects. I = do think they'll need to
leverage the economy of scale driven by ot= hers. An ASIC tape out, just
for the design, is ~$80M and a minimum of = 18 mos of high-skill,
engineering work by many specialties, signal inte= grity, etc. Then, after
all that, one has to get in line with a foundry= that needs to produce in
volume per their mfg economies of scale. Thes= e markets fundamentally
have to be driven by large orders from provider= s with millions of
subscribers. That's just the market & engine= ering reality of things.


Every ASIC necessary to deploy is already on the market = in high volume.=C2=A0 No additional ~$80M needs spent. =C2=A0~$80M that MUS= T come from the customer at the end of the day.=C2=A0 Another increase in b= roadband costs. =C2=A0 Every massive change you suggest will pull money fro= m actually running mainline fiber to communities where various technologies= can already deliver huge speeds at low latency.=C2=A0 I=E2=80=99m an opera= tor, my primary limitations logistically speaking is inability to get 10Gbp= s+ fiber off the existing fiber footprint.=C2=A0 Even the lowly DSL footpri= nt could be upgraded with relative ease to get a few hundred Mbps if, and f= orgive me for leaning not his so hard, the previously designed monopoly tha= t owned not only the copper plant but also the fiber that is already there = wasn=E2=80=99t waiting around for the next government hand out before upgra= ding. =C2=A0 Fiber to the DSLAM and VSDL would be a nearly instant upgrade = to 100+ x 50+ speeds for easily 80% of rural users. =C2=A0

We don=E2=80=99t need a completely different model (FiWi) when we have a= ll of the parts and pieces in mass production and available right now, we h= ave a political system that promotes monopoly and actively encourages them = to wait until either a self funded competitor moves in or government money = shows up with mandates.=C2=A0 There is no reason at all to have 3-7Mbps DSL= in most of America.=C2=A0 This is not a technical limit.=C2=A0

An aspect of the FiWi argument is that = these NRE spends today and
tomorrow are mostly from SERDES & lasers= /optics in the data centers and
the CMOS radios & PHYs in handsets.= Let us look here for the thousands
of engineers needed and for the sup= ply of parts for the next decade+. I
don't see it coming from anywh= ere else.
We have 100G hardware routers from mu= ltiple vendors, Qualcomm, Broadcom, Marvell.=C2=A0 We have 1-100G optics on= the market today for cheap.=C2=A0 Marvell makes a line of chips that can d= o 40Gbps hardware switch or routed for like $20, get=E2=80=99s put in $200 = MikroTik devices today. A grand gets you into a device that can do 100G tod= ay.=C2=A0 Obviously that=E2=80=99s from the cheapest vendor but 2-10x that = price will get you into the =E2=80=98good stuff=E2=80=99.=C2=A0 We already = have this.


Then we need = the in-premise fiber installers and the OSP labor forces
who are critic= al to our success.

And finally, it's the operations & manage= ment and the reduction of those
expenses in a manner that scales.

Where exactly are the costs, operations, and management savings here?

Basically this leads me to the question which I=E2=80=99m ask= ing with an attempt to avoid condescension, do you/have you run an ISP?=C2= =A0 My operations and management costs are primarily customer service and l= ogistic (vehicles, labor, and so on) and not network management.

Fiber in-premise has a negative value.=C2=A0 It=E2=80=99s more exp= ensive to terminate and repair, port costs are more, vastly (like 100x) mor= e likely to damage a fiber patch cable vs cat5e, and the advantages of fibe= r are lost on short distances. =C2=A01,2.5, 5, and 10G copper is easy, chea= p to terminate, cheap to install, cheap ports in switches, cheap ports on d= evices, and fast.=C2=A0 The entire =E2=80=98need=E2=80=99 for fiber in this= context is the FiWi concept of centralized networking which again IMO is s= omething ALL IT/MSP will outright reject killing it off for business uses a= nd will not fare well for consumers who are concerned more and more about p= rivacy.

Just my opinion here, but the entirety of the FiW= i concept will be dead on arrival with almost all opposing it and only a fe= w supporters.
--00000000000002ae6805f80f1a65--