From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.lang.hm (unknown [66.167.227.145]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 8926B3B29D for ; Fri, 4 Mar 2022 13:14:33 -0500 (EST) Received: from dlang-mobile (unknown [10.2.2.69]) by mail.lang.hm (Postfix) with ESMTP id 5D08D1247A8; Fri, 4 Mar 2022 10:14:32 -0800 (PST) Date: Fri, 4 Mar 2022 10:14:32 -0800 (PST) From: David Lang To: Ulrich Speidel cc: David Lang , Mike Puchol , starlink@lists.bufferbloat.net In-Reply-To: <7ee92a7f-ae58-5090-8ee0-32df8ec29c2b@auckland.ac.nz> Message-ID: <712r509p-7on6-5657-3172-n9140n9097@ynat.uz> References: <1646351242.121623495@apps.rackspace.com> <2b2f1808-8cf7-6eae-3157-a5fc554a2424@auckland.ac.nz> <03e73122-63c3-497b-82c6-b7b7f23b627a@Spark> <3ooo342q-s937-qq3-492q-723np793qoo@ynat.uz> <7ee92a7f-ae58-5090-8ee0-32df8ec29c2b@auckland.ac.nz> MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset=US-ASCII Subject: Re: [Starlink] Starlink Digest, Vol 12, Issue 6 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: Fri, 04 Mar 2022 18:14:33 -0000 On Fri, 4 Mar 2022, Ulrich Speidel wrote: > In terms of Starlink - I really think that it's a red herring, at least for > now. As I said, if Starlink can't muster anywhere near enough satellite > capacity to serve all of a small town in Montana that's surrounded by > gateways close by, then it's not going to be replacing the Internet as we > know it in a country 60% larger in area and 40 x larger in population. At > best it might be able to provide some backup in a relatively small number of > places. It depends on what you set as your requirements. If you are talking about everyone streaming video, you are correct, but if you talk about less bandwidth intensive uses, a little bandwidth goes a long ways. There's also FAR more of a difference between nothing and low bandwidth than between low and high bandwidth. Telephone audio is an 64Mb stream, without compression, email and text chat are very low bandwidth. 20 years ago, you could have an office of 100 employees living on a 1.5Mb Internet connection and have people very happy. A single dishy is 100x this. I agree that Starlink is not a full replacement for hard-wired Internet, and it never will be. But the ability to get that much bandwidth into an area that doesn't have wired Internet wihtout requiring special crews to come in and set up the infrastructure (like you would for geostationary dishes) is a huge step forward for disaster relief. With capabilities like this now available, we (the tech community) need to look at options to be able to extend this connectivity from a point source across a wider area (ways to do mesh and have it not collapse, understanding channnel allocations, sane directional antenna uses, etc) including how to provide power. And also take a careful look at the bandwith that apps are using and find ones that are sane to use. Since (almost) everyone has phones as endpoints now, having the ability to put a voip app on the phones and have them able to call and text chat freely within the connectivity bubble without any need to use the external bandwith, but be able to connect out in a fairly transparent manner (think how long distance calls were something significant 40-50 years ago, but were still using the same equipment and basic process). Can such apps indicate to the user if they are talking to someone really local (say sharing the same wifi), or more remote, so that they can How can such apps be made available to the people with phones? (Apple makes it really hard to side-load apps for example), How can the services get bundled (raspverry pi or live CD linux images that provide these services and the app images to download for example). What can be done with OpenWRT builds to make turnkey conversions of APs into bandwidth-efficient mesh nodes. This includes how a bit of wire can go a long way towards making a wifi system work better. How can we bundle lessons for techies on the ground to teach them what to do (and what not to do) in setting these things up? David Lang