Re: V6 still not supported
On 24Mar22, Pascal Thubert (pthubert) allegedly wrote: > Hello Mark: > > > Any such "transition plan" whether "working" or "straightforward" is > > logically impossible. Why anyone thinks such a mythical plan might yet be > > formulated some 20+ years after deploying any of ipv6, ipv4++ or ipv6-lite > > is > > absurd. > > This is dishonest My point is that if there was a real transition plan it would have been invented and executed by now and we'd all be on ipv6. Yet the reality is that here we are some 20 years later with no plan and no ubiquitous ipv6. How is that observation dishonest? > considering that I just proved on this very thread that such ideas existed I don't know why you're conflating an idea with a plan - they are about as far away from each other as is possible. Frankly no one cares about ideas, they're a dime a dozen. A plan is an actionable, compelling and logical set of steps towards an end result. Do you have such a thing for moving everyone on the planet to ipv6? Here's a simple test of whether you have a plan or not. I'm connected via my legacy ipv4 ISP router completely oblivous to ipv6. How does your plan incentivise me to upgrade my router to support ipv6? When you have an answer to that, you might have a glimmer of a plan. Mark.
Re: V6 still not supported
On 24Mar22, Vasilenko Eduard allegedly wrote: > Hence, the primary blocking entity for IPv6 adoption is Google: they do not > support DHCPv6 for the most popular OS. No. The primary "blocking entity" is that "legacy" ipv4 works just fine and adopting ipv6 or ipv4++ or ipv6-lite or ipv-magical is harder than doing nothing. That Google/Android don't like DHCPv6 is largely irrelevant. My five year old ISP router only supports ipv4. Yet I get to every site on the planet just fine. Give me one good reason to spend my hard earned pay on an expensive new router which supports ipv6? You have two choices: make my ipv4 router fail or make the new ipv6 router compelling. How do you propose to do either of those? Mark.
Re: V6 still not supported
On 24Mar22, Greg Skinner via NANOG allegedly wrote: > straightforward transition plan > in-hand working transition strategy > nor a straightforward transition Any such "transition plan" whether "working" or "straightforward" is logically impossible. Why anyone thinks such a mythical plan might yet be formulated some 20+ years after deploying any of ipv6, ipv4++ or ipv6-lite is absurd. The logic goes: we support legacy "do nothing" ipv4 deployments forever. We also expect those same deployments to invest significant effort, cost and risk to move off their perfectly functioning network for no self-serving benefit. There be unicorns and denial of human nature. Mark.
Re: V6 still not supported
On 23Mar22, Owen DeLong via NANOG allegedly wrote: > I would not say that IPv6 has been and continues to be a failure Even if one might ask that question, what are the realistic alternatives? 1. Drop ipv6 and replace it with ipv4++ or ipv6-lite or whatever other protocol that magically creates a better and quicker transition? 2. Drop ipv6 and extend above the network layer for the forseeable future? By extend I mean things which only introduce ipv4-compatible changes: NATs, TURN, CDN at the edge, application overlays and other higher layer solutions. 3. Live with ipv6 and continue to engineer simpler, better, easier and no-brainer deployments? I'll admit it risks being a "sunk cost falacy" argument to perpetuate ipv6, but are the alternatives so clear that we're really ready to drop ipv6? > so much as IPv6 has not yet achieved its goal. As someone previously mentioned, "legacy" support can have an extremely long tail which might superficially make "achieving a goal" look like a failure. Forget ss7 and SIP, what about 100LL vs unleaded petrol or 1/2" bolts vs 13mm bolts? Both must be 50 years in the making with many more years to come. The glacial grind of displacing legacy tech is hardly unique to network protocols. In the grand scheme of things, the goal of replacing ipv4 with ipv6 has really only had a relatively short life-time compared to many other tech transitions. Perhaps it's time to adopt the patience of the physical world? Mark.
Re: V6 still not supported
On 19Mar22, Matt Hoppes allegedly wrote: > So, while it's true that a 192.168.0.1 computer couldn't connect to a > 43.23.0.0.12.168.0.1 computer, without a software patch - that patch > would be very simple and quick to deploy Let's call this ipv4++ Question: How does 192.168.0.1 learn about 43.23.0.0.12.168.0.1? Is that a DNS lookup? How does the DNS support ipv4++ addresses? Is that some extension to the A RR? It better be an extension that doesn't break packet validation rules embeded in most DNS libraries and middleware. You give 'em an A RData longer than 32 bits and they're going to drop it with prejudice. Perhaps you should invent a new ipv4++ address RR to avoid some of these issues? In either case, how does every program on my ipv4 computer deal with these new addresses that come back from a DNS lookup? Do you intend to modify every DNS library to hide these extensions from older programs? How do you do that exactly? What about my home-grown DNS library? Who patches that? Here's a code fragment from my ipv4-only web browser: uint32 ip ip = dnslookup("www.rivervalleyinternet.net", TypeA) socket = connect(ip) What does 'ip' contain if www.rivervalleyinternet.net is ipv4++ compliant and advertises 43.23.0.0.199.34.228.100? Do these magical concentrators sniff out DNS queries and do some form of translation? How does that work with DoH and DoT? Or are you suggesting that www.rivervalleyinternet.net continues to advertise and listen on both 43.23.0.0.199.34.228.100 *and* good ol' 199.34.228.100 until virtually every client and network on the planet transitions to ipv4++? In short, the transition plan is to have www.rivervalleyinternet.net run dual-stacked for many years to come. Yes? Speaking of DNS lookups. If my ipv4++ DNS server is on the same LAN as my laptop, how do I talk to it? You can't ARP for ipv4++ addresses, so you'll have to invent a new ARP protocol type or a new LAN protocol. Is that in your patch too? Make sure the patch applies to network devices running proxy ARP as well, ok? If I connect to an ipv4++ network, how do I acquire my ipv4++ address? If it's DHCP, doesn't that require an extension to the DHCP protocol to support the larger ipv4++ addresses? So DHCP protocol extensions and changes to all DHCP servers and clients are part of the patch too, right? Or perhaps you plan to invent a new DHCP packet which better accomodates all of the ipv4++ addresses that can get returned? Still plenty of code changes. And how do I even do DHCP? Does ipv4++ support broadcast in the same way as ipv4? What of DHCP relays? They will need to be upgraded I presume. So let's say we've solved all the issues with getting on a network, talking over a LAN, acquiring an ipv4++ address, finding our ipv4++ capable router and resolving ipv4++ addresses. My application is ready to send an ipv4++ packet to a remote destionation. But what does an ipv4++ packet look like on the wire? Is it an ipv4 packet with bigger address fields? An ipv4 packet with an extension? Or do you propose inventing a new IP type? Do these packets pass thru ipv4-only routers untainted or must they be "concentrated" beforehand? Won't all the firewalls and router vendors need to change their products to allow such packets to pass? Normally oddball ipv4 packets are dropped of course. As we know, vendor changes can notoriously take decades; just ask the ipv6 crowd. Ok. We've upgraded all our infrastructure to route ipv4++ packets. The packet reached the edge of our network. But how does the edge know that the next hop (our ISP) supports ipv4++ packets? Do routers have to exchange capabilities with each other? How do they do that? For that matter, how does my application know that the whole path thru to the destination supports ipv4++? It only needs one transition across an ipv4-only network somewhere in the path and the packet will be dropped. I think you're going to have to advertise ipv4++ reachability on a per-network basis. Perhaps using BGP? Oh, so that means your patch has to add ipv4++ capabilities to all BGP implementations. Hmm. > However, to get back 192.168.0.1 can proxy through an IPv4.1 to IPv4.2 > concentration system. Sounds like you're proposing a global deployment of reverse CG-NATs (aka concentrators) which need to run for as long as the transition takes. Do you run concentrators on ISP borders where one party is ipv4 and the other party is ipv4++? Not sure a lot of edge/border routers have the memory footprint or CPU grunt to maintain NAT tables that might involve 1,000s or millions of flows. Your patch might have to include a memory upgrade. Will these address ranges fit into 32bit CAM or will routers need to drop back to regular memory and software lookups? That's a pretty nasty performance hit which was an argument against ipv6 for quite a while. One of the bigger problems with your "concentrators" (apart from who implements t
Re: Authoritative Resources for Public DNS Pinging
On 09Feb22, Joe Greco allegedly wrote: > I dunno. I think I'd find that being unable to resolve a hostname or > being unable to exchange packets result in a similar level of Internet > brokenness. Sure. The result is the same, but as a discriminator for diagnosing the problem it's quite different. If a support rep hears "cannot reach fatbook" vs. ping 8.8.8.8 fails they'll hopefully go down a different diagnostic route. In large partI think of this sort of tool as a "what next" helper or "where do I look next" helper. > It is going to be hard to quantify all the things you might > want to test for. You already enumerated several. But if it has to be > a comprehensive "Internet is fully working" test, what do you do to be > able to detect that your local coffee shop isn't implementing a net nanny > filter? Just to take it too far down the road. ;-) Well comprehensive might be a bit much, but useful info for a support rep or a tech forum helper, well, that's a different matter. The Ocean boiling we aint. In many ways the idea is simply to automate the basic steps that most of us would take to diagnose a connectivity issue: for ip in 4, 6 1a) Can I reach first hop - check 1b) Can I reach ISP first hop - check 2a) Can I reach a resolver - check 3a) Can I resolve ISP end-points - check 3b) Can I reach ISP end-points - check 4a) Can I resolve ping.ripe.net - check 4b) Can I reach ping.ripe.net end-points - check done Or similar. The tool bootstraps its way up rather than offering just a binary yeah/nay. The output might simply be the above. I reckon if you saw that sort of output you'd be able to make a pretty informed guess as to where the problem might be - or at least what to do next. Which is the goal. We know we have a problem, we want to zero in on it. >From an automation POV it's nothing more than a bit of data-gathering and >issuing pretty standard network exchanges in sequence. By way of example, on FreeBSD13, ping is 4,000+ lines of C. I'm certain the above could be implemented in far fewer ELOCs in a modern language. And, if these basic steps are augmented by ISPs adopting one or two conventions such as well-defined DNS names and end-points, then such a tool could offer a lot of insights for that inevitable support call: "I ran 'ping-internet [sol.net]' and it said: ...". As J. Hellenthal mentioned earlier, I think there's an argument here for why ISPs might encourage such a beast. What do you think? Mark.
Re: Authoritative Resources for Public DNS Pinging
On 09Feb22, Joe Greco allegedly wrote: > So what people really want is to be able to "ping internet" and so far > the easiest thing people have been able to find is "ping 8.8.8.8" or > some other easily remembered thing. Yes, I think "ping internet" is the most accurate description thus far. Or perhaps "reach internet". > Does this mean that perhaps we should seriously consider having some > TLD being named "internet" Meaning you need to have a functioning DNS resolver first? I'm sure you see the problem with that clouding the results of a diagnostic test. > service providers register appropriate upstream targets for their > customers, and then maybe also allow for some form of registration such > that if I wanted to provide a remote ping target for AS14536, I could > somehow register "as14536.internet" or "solnet.internet"? Possibly. You'd want to be crystal clear on the use cases. As a starting point, maybe: 1. Do packets leave my network? 2. Do packets leave my ISP's network? 3. Mainly for IOT - is the internet reachable? Because of 2 and 3. I don't think creative solutions such as ISPs any-casting some memorable IP or name will do the trick. And because of 1. anything relying on DNS resolution is probably a non-starter. Much as I like "ping ping.ripe.net" it alone is too intertwined with DNS resolution to be a reliable alternative. > Fundamentally, this is a valid issue. Yup. There are far more home-gamers and tiny network admins (the networks are tiny, not the admins) who just want to run a reachability test or add a command to a cheap network monitor cron job. Those on this list who can - or should - do something more sophisticated are numerically in the minority of people who care about reachability and are not really the target audience for a better "ping 8.8.8.8". > and we'll end up needing a special non-ping client and some trainwreck of > names and > other hard-to-grok I'm not sure the two are fundamentally intertwined tho it could easily be an unintended consequence. However, being constrained to creating a new ping target does severely limit the choices. And including ipv6 just makes that more complicated. The other matter is that the alternative probably has to present a compelling case to cause change in behavior. I can see an industry standard ping target being of possible use to tests built into devices. But again it'd have to be compelling for most manufacturers to even notice. But for humans, I'd be surprised if you can create a compelling alternative ping target. For them, I'd be going down the path of a "ping-internet" command which answers use-cases 1. & 2. while carefully avoiding the second-system syndrome - he says with a laugh. Mark.
Re: Authoritative Resources for Public DNS Pinging
On 08Feb22, Mike Hammett allegedly wrote: > Some people need a clue by four and I'm looking to build my collection of > them. > "Google services, including Google Public DNS, are not designed as ICMP > network testing services" Hard to disagree with "their network, their rules", but we're talking about an entrenched, pervasive, Internet-wide behaviorial issue. My guess is that making ping/ICMP less reliable to the extent that it becomes unusable wont change fundamental behavior. Rather, it'll make said "pingers" reach for another tool that does more or less the same thing with more or less as little extra effort as possible on their part. And what might such an alternate tool do? My guess is one which SYN/ACKs various popular TCP ports (say 22, 25, 80, 443) and maybe sends a well-formed UDP packet to a few popular DNS ports (say 53 and 119). Let's call this command "nmap -sn" with a few tweaks, shall we? After all, it's no big deal to the pinger if their reachability command now exchanges 10-12 packets with resource intensive destination ports instead of a couple of packets to lightweight destinations. I'll bet most pingers will neither know nor care, especially if their next-gen ping works more consistently than the old one. So. Question. Will making ping/ICMP mostly useless for home-gamers and lazy network admins change internet behaviour for the better? Or will it have unintended consequences such as an evolutionary adaptation by the tools resulting in yet more unwanted traffic which is even harder to eliminate? Mark.
Re: NDAA passed: Internet and Online Streaming Services Emergency Alert Study
On 03Jan21, Brandon Martin allegedly wrote: > I was thinking more in the original context of this thread w.r.t. > potential distribution of emergency alerts. That could, if > semi-centralized, easily result in 100s of million connections to juggle > across a single service just for the USA. While it presumably wouldn't > be quite that centralized, it's a sizable problem to manage. Indeed. But how do you know the clients are still connected? And if they aren't, there is not much a server can do beyond discarding the state. Presumably the client would need to run a fairly frequent keep-a-live/reconnect strategy to ensure the connection is still functioning. Which raises the question: how long a delay do you tolerate for an emergency alert? I think the end result is a lot of active connections and keep-a-live traffic. Not really quiescent at all. In the end, probably just as cheap to poll a CDN. Mark.
Re: NDAA passed: Internet and Online Streaming Services Emergency Alert Study
On 03Jan21, Brandon Martin allegedly wrote: > On 1/3/21 3:11 PM, Jay R. Ashworth wrote: > > Well, TCP means that the servers have to expect to have 100k's of open > > connections; I remember that used to be a problem. > > Out of curiosity, has anyone investigated if it's possible to hold open > a low-traffic, long-lived TCP session without actually storing state > using techniques similar to syncookies and do so in a compatible manner? Creating quiescent sockets has certainly been discussed in the context of RSS where you might want to server-notify a large number of long-held client connections very infrequently. While a kernel could quiesce a TCP socket down to maybe 100 bytes or so (endpoint tuples, sequence numbers, window sizes and a few other odds and sods), a big residual cost is application state - in particular TLS state. Even with a participating application, quiescing in-memory state to something less than, say, 1KB is probably hard but might be doable with a participating TLS library. If so, a million quiescent connections could conceivably be stashed in a coupla GB of memory. And of course if you're prepared to wear a disk read to recover quiescent state, your in-memory cost could be less than 100 bytes allowing many millions of quiescent connections per server. Having said all that, as far as I understand it, none of the DNS-over-TCP systems imply centralization, that's just how a few applications have chosen to deploy. We deploy DOH to a private self-managed server pool which consume a measly 10-20 concurrent TCP sessions. Mark.
Re: An appeal for more bandwidth to the Internet Archive
On 13May20, Denys Fedoryshchenko allegedly wrote: > What about introducing some cache offloading, like CDN doing? (Google, > Facebook, Netflix, Akamai, etc) > Maybe some opensource communities can help as well Surely someone has already thought thru the idea of a community CDN? Perhaps along the lines of pool.ntp.org? What became of that discussion? Maybe a TOR network could be repurposed to cover the same ground. Mark.
Re: alternative to voip gateways
> We need to keep battery backup requirements, and expand them to all last > mile IP bits. The need to call 911 has not gone away. For sure. I was merely observing that the conversion of POTS to VOIP in Australia didn't create a nation-wide disaster as the pearl-clutchers once predicted. In fact, if anything, the same folk who complained about the so-called largesse of a nationwide IP last-mile are strangely silent now that WFH is de rigueur. To your point, the original plan was 90+% passive optical back to major exchanges so the infrastructure was largely invulnerable to wide-scale power shutdowns/failures. All a residence has to do is feed a 7W NTD to stay connected. Mark.
Re: alternative to voip gateways
> wasnt there a hige shit stom in australia for their new national > broadband network making internet ptrimary and phone secondary, a lot > of aussies on forums I frequent bitch about its reliability, where > even their aged copper services worked fine, not to mention prolonged > outages due to storms and the bushfires they had recently, lets hope > the world learns from australias mistakes and not go down that path. There are still a few complaints every now and again but fixed line numbers are continuing to drop off a cliff and those residential services which remain have almost all been converted to VOIP via home gateway with an FXS port. Mobile/Cell is where most people ended up. Especially since you can get unlimited calls/txt with some data for about ten bucks a month. It also helps that a number of the mobile providers include wifi-calling so mobile is a viable alternative even in weak cell coverage areas if you have internet. Yes, everyone knows about the reliability/power-failure arguments but in the latest set of bushfires whole exchanges, backhaul services and power distribution cables were destroyed so 8 hours of battery backed up POTS in a local exchange didn't help much. Mark.
Re: Elephant in the room - Akamai
> Have there been any fundamental change in their network architecture > that might explain pulling these caches? Maybe not network architecture, but what if the cache-to-content ratio is dropping dramatically due to changes in consumer behavior and/or a huge increase in the underlying content (such as adoption of higher and multiple-resolution videos)? There has to be a tipping point at which a proportionally small cache becomes almost worthless from a traffic saving perspective. If you run a cluster one presumes you can see what your in/out ratio looks like and where the trend-line is headed. Another possibility might be security. It may be that they need additional security credentials for newer services which they are reluctant to load into remote cache clusters they don't physically control. Mark.
Re: all major US carriers received text messages overnight that appear to have been sent around Valentine's Day 2019
> What I've seen happen more often than that: > > Server goes partly belly-up, queue fills up. Backup process runs, backing up > the > queue. (Optionally here: Reboot the server and lose the queue). Much later, > the > server hits another issue that requires recovering from backups - and they > restore > a truly ancient copy. Particularly as mail servers tend to check for expired messages *after* a delivery attempt. This means a restored queue containing ancient messages will most likely be given one last delivery attempt prior to bouncing. One real example comes from the qmail-send man page: queuelifetime Number of seconds a message can stay in the queue. Default: 604800 (one week). After this time expires, qmail-send will try the message once more, but it will treat any temporary delivery failures as permanent failures. Combine that with the fact that it's not unheard of for SMS servers to be derived from mail servers (since they do virtually the same thing) and an accidental queue restore or server revivication seems the most plausible. Mark.
Re: IPv6 on mobile networks, was Update to BCP-38?
> Yep I see this on AT&t's post paid network with my Pixel 3A XL as well, one > place I really noticed it causing issues is with Facebook and Instagram > where Facebook requires constant captions to view any Facebook links I > receive and embedded Instagram content in news articles and things of that > nature often failed load. It is very annoying. > > Tmobile US, VZ, and Sprint all have IPv6, but only AT&T has this behavior > > afaik. HTTP proxies are used by some mobile carriers to down-scale media sent thru their radio network to reduce bandwidth. They rationalise that, e.g, a HD video can be down-scaled for a tiny screen with no real loss of fidelity but a signficant reduction in bandwidth. Similar strategies apply to almost all compressible media: mp3, jpegs, etc. More often used outside the US as I recall but sounds like AT&T might be doing something similar. You could try a mobile fetch of a known media file via HTTP and HTTPS then compare them for possible insights (make sure to use a mobile browser to avoid browser-detects). Such proxies are sometimes used for carrier ad-insertion as well so one presumes they detest the widespread switch to HTTPS for at least two reasons. Mark.