Show Idle (>14 d.) Chans
← 2021-11-20 | 2021-11-22 → ↓
dulapbot: Logged on 2021-11-17 10:47:10 signpost: to wish to have been there is to wish to fail with a lost cause. better to know why they failed and fix it.
crtdaydreams: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066394 << If you know other hackers my age, please introduce me.
dulapbot: Logged on 2021-11-17 12:04:00 BingoBoingo: And generally the young adults today seem less retarded than self and peers at their age, but maybe I'm just picky about the ones I entertain?
crtdaydreams: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066435 << I've been reading a lot of your articles over the past few days, my technical knowledge provides a huge limitation on the intrinsic understanding concepts like C-gates and dataflow computing.
dulapbot: Logged on 2021-11-17 15:08:10 asciilifeform: ( and could return en masse, if a sane iron were to be fabbed ! )
crtdaydreams: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066480 << I will take that into careful consideration moving forward.
dulapbot: Logged on 2021-11-17 15:32:09 signpost: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066436 << many of the precocious fail later because narcissism deranges
crtdaydreams: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066494 << Any top 10 tips on how *not* to be a narcissist?
dulapbot: Logged on 2021-11-17 16:12:35 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066480 << possibly not merely many but most
crtdaydreams: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066582 << /me hasn't actually bought any crypto after dumping shitcoin a few days prior.
dulapbot: Logged on 2021-11-17 22:43:08 vex: did you manange to swap your bullshit out for something useful crtdaydreams?
dulapbot: Logged on 2021-11-18 00:13:53 vex: I am super keen to test certain newbies arthmetic skills
dulapbot: Logged on 2021-11-18 11:29:50 asciilifeform: bistable -- 1 collapses into the other periodically.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-21#1066949 << imho it may be a birth defect. (sometimes latent, like schizos, may need a 'trigger')
dulapbot: Logged on 2021-11-21 05:02:34 crtdaydreams: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-17#1066494 << Any top 10 tips on how *not* to be a narcissist?
asciilifeform: $ticker btc usd
busybot: Current BTC price in USD: $59433.46
asciilifeform: !w poll
watchglass: Polling 17 nodes...
watchglass: 176.9.59.199:8333 : Could not connect! (Operator: jurov)
watchglass: 185.85.38.54:8333 : Could not connect!
watchglass: 84.16.46.130:8333 : Could not connect!
watchglass: 205.134.172.26:8333 : Alive: (0.081s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=710711
watchglass: 71.191.220.241:8333 : (pool-71-191-220-241.washdc.fios.verizon.net) Alive: (0.099s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=710711 (Operator: asciilifeform)
watchglass: 185.163.46.29:8333 : Could not connect!
watchglass: 213.109.238.156:8333 : Could not connect!
watchglass: 54.39.156.171:8333 : (ns562940.ip-54-39-156.net) Alive: (0.111s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=710711
watchglass: 205.134.172.27:8333 : Alive: (0.083s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=710711 (Operator: asciilifeform)
watchglass: 205.134.172.6:8333 : (172-6.core.ai.net) Alive: (0.141s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=710711
watchglass: 205.134.172.4:8333 : (172-4.core.ai.net) Alive: (0.142s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=710711
watchglass: 205.134.172.28:8333 : Alive: (0.084s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=710711 (Operator: whaack)
watchglass: 208.94.240.42:8333 : Alive: (0.169s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=710711
watchglass: 143.202.160.10:8333 : Alive: (0.194s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=710711
watchglass: 54.38.94.63:8333 : (ns3140226.ip-54-38-94.eu) Alive: (0.258s) V=88888 (/therealbitcoin.org:0.8.88.88/) Jumpers=0x1 (TRB-Compat.) Blocks=710711
watchglass: 103.36.92.112:8333 : Violated BTC Protocol: Bad header length!
watchglass: 192.151.158.26:8333 : Busy? (No answer in 100 sec.)
asciilifeform: continuing this thrd : asciilifeform realized that he neglected to specify what happens when ~no one~ supplies the missing msg within the set time interval. (possibly add to a queue for periodic requests? but when to 'give up' ? can't keep these around 4evah) -- comments invited.
dulapbot: Logged on 2021-11-20 17:23:52 asciilifeform: i.e. bitwise agreement, 'we're all sharing wikilicks.tar.gz and hash of it is H'. rather than 'we each have a log history with possibly 10% missing, each in different spot'
asciilifeform: ^ thimbronion et al
asciilifeform: anuther observation -- a station possib. should have some awareness of whether an originated message actually went out before updating own net/self hash. (we dun have 'acks', but a broadcast msg is expected to come back verbatim, given as rebroadcasts are to all peers incl. originator)
asciilifeform: i.e. if you're transmitting but yer cable is unplugged, possibly pest oughta warn, rather than potentially pissing out the msg n days later when you plug back in
jonsykkel: i dont understand why dedup window is larger than 30min timestump window
asciilifeform: jonsykkel: station clocks are permitted to differ by +/- 15m. this, plus margin.
punkman: I don't have a dedup window in my implementation, only timestamp window
punkman: meaning, anything that passes timestamp window, will be checked against all messages I have on disk
punkman: muh scalabilities "One 256 bit integer costs 20,000 gas to save, which at today’s costs is about $7.50… for one number!"
punkman: how do you even trade shittokens at those prices, I've yet to figure out
signpost: I believe the notion behind having a window is fixed bounds on dedup lookup
signpost: rather than ever-growing
signpost: "all things shall be worst case and worst case fixed"
jonsykkel: asciilifeform: will not worst case be: time() == 100min. receive valid paket A stumped 115min. put in dedup. 30mins later at time() == 130min someone replays A. still passes stump window. so has to be deduped
asciilifeform: signpost: correct
jonsykkel: punkman: cannot someone then assrape ur io by spamming msgs inside stump window?
asciilifeform: jonsykkel: plz elaborate (what means 'time() == 100min' ?)
jonsykkel: asciilifeform: ur own station clock returns 100min
asciilifeform: jonsykkel: timestamps are unix epochal times (64bit)
jonsykkel: yes i mean it can be any offset
punkman: I have an index of msg hashes, if it grows too big I suppose there might be a performance hit. but it's minor optimization at this point
asciilifeform: jonsykkel: the expiration interval is 15min. i.e. nuffin you accept at time t is valid at t+15min.
asciilifeform: jonsykkel: or do i misunderstand your q ?
asciilifeform: punkman: the notion is that a reasonable iron (e.g. 32core 'dulap') oughta be able to process incoming packets at Gb/s
asciilifeform: punkman: this is impractical if an opponent can force you to access disk to do it
jonsykkel: asciilifeform: ill fix my text
asciilifeform: ( see also re subj )
punkman: will be interesting to run some kind of benchmark when I get it back to working
asciilifeform: jonsykkel: keep in mind that the only type of spamola that costs palpable cpu is replays of ~valid~ packets (i.e. from a peer) -- and, of course, packets actually generated by a peer (the solution to this is arguably a 'hog' alarm + unpeer)
jonsykkel: asciilifeform: will not worst case be: time() == t. receive valid paket A stumped t+15min (which is <= t+15min so it passes stump window). put in dedup. 30mins later at time() == t+30min someone replays A. still passes stump window (cuz A.stump == t+15min, which is >= t+30min-15min). so has to be deduped
asciilifeform: s/stumped/stamped neh
jonsykkel: asciilifeform: yes, my question is why the dedup bufer needs to store 1hour worth of hashes+stumps rather than 30min
asciilifeform: and correct, if station received a current+15m packet, it can be replayed 30m later. hence why dedup queue holds 1h min.
asciilifeform: jonsykkel: margin.
asciilifeform: jonsykkel: y'know, how when building bridge and expecting 10 tanks to fit, you build for e.g. 25.
jonsykkel: alrite
jonsykkel: yeah, i guess i cant imagine what exactly would cause margin of 10sec to be insufficient
asciilifeform not aware of any extant irons where clock can drift by 10s of min within span of hour, but theoretically possible, for instance.
asciilifeform: 1 simple example -- you realized that yer clock was +15m. and fixed.
asciilifeform: this must not disrupt pest station.
jonsykkel: well if your clokc is +15m it might as well be +16m, in which case you have bigger problems
asciilifeform: jonsykkel: says right there in spec, that operators are expected to keep within 15m.
asciilifeform: and that this is achievable w/out use of automated ntpisms. which is historically correct.
jonsykkel: this is what i mean, if your clock is +16min relative to peers, all msgs will be rejected so you have bigger problems than dedup margin
asciilifeform: jonsykkel: correct. but operator must be able to set his +15min clock back -15min w/out disrupting station.
jonsykkel: if operator sets to +15min and his peer sets to -15min ur fukd tho
asciilifeform: hence '+/- 15m of each other' verbiage in spec.
asciilifeform: this is the responsibility of the operators, and there's no pill against this.
jonsykkel: ah misread ur last msg
asciilifeform: i.e. if you can't be arsed to keep within +/-15m of other people, you can't pest.
jonsykkel: sure, agreed
asciilifeform: jonsykkel: dunno if you were here for the ancient threads re p2p, but 1 key objective is to make protocol that can be readily hardwarized, i.e. can build box which eats packets from street at e.g. 1Gb/s in one hole, and passes only valid packets out of other hole (and vice versa)
asciilifeform: more interestingly, oughta be able to connect N such boxes (each on diff. ip, or even diff. isp) and connect'em all to 1 station
asciilifeform: i.e. to make ddos as close to physically impossible as possible.
asciilifeform: ( see also . )
dulapbot: Logged on 2021-09-12 13:31:32 asciilifeform: asciilifeform's scheme does not try to do the impossible (e.g. opposable signatures at line rate on pc) but next best thing -- nomoar ddos (reject at Gb/s easily on even 8core box; agnostic of ip, so if you have 9000 nics -- you're in biznis for so long as ~1~ can send/receive packets; rejection of replays; and finally, you get to have a ~reasonable~ if not perfect idea that you're talking at time T to same person as a
asciilifeform: i.e. if you have N nics, connected to N isps, while ~one~ of'em is able to send/receive packets, your station remains in biz.
asciilifeform: (and, importantly, the enemy has no ready means of knowing what N is, because the spares aint in use unless req'd)
asciilifeform: potentially 1 or moar of the links could even be a private line (radio, hell, dialup, whatever)
asciilifeform: jonsykkel: 'secret' objective of pest is to build a wotronic internet parasitically on top of existing comms.
jonsykkel: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-21#1067034 << what are the boxes here? idea is that you would do verification and deduping steps on separate hardware?
dulapbot: Logged on 2021-11-21 13:26:14 asciilifeform: jonsykkel: dunno if you were here for the ancient threads re p2p, but 1 key objective is to make protocol that can be readily hardwarized, i.e. can build box which eats packets from street at e.g. 1Gb/s in one hole, and passes only valid packets out of other hole (and vice versa)
asciilifeform: jonsykkel: nah. simply 1st pass filter (reject liquishit)
jonsykkel: right, so it spits out deliquishited black packets out of other hole
jonsykkel: makes sense
jonsykkel: yeah parasitic internet is good idea, cant wait
asciilifeform: jonsykkel: there's 2 basic effects of ddos. 1 is to simply fill up your line, and effect is similar to a cut cable (you have a 1Gb/s fiber, over9000 bots are sending you rubbish, for total of 100Gb/s, isp's filter kicks in and you have a however long outage on the line.) the other effect is to eat cpu/disk on the victim's end
asciilifeform: the pill against the former is simply to be able to operate over N fallback links, w/out enemy necessarily having any idea of their existence. the pill against the latter is to actually have a protocol where a spamola packet is rejected no later than 1/t, where t is the interval the nic takes to receive the packet.
asciilifeform: afaik pest is the only published protocol which explicitly mandates both.
asciilifeform: err s/1\/t/t
asciilifeform: jonsykkel: makes sense ?
jonsykkel: right right
jonsykkel: slow brain takes a while to process
asciilifeform: jonsykkel: dunno how familiar you are with mechanics of ddos; but e.g. tcp makes it quite simple (once somebody gives you a SYN, you're stuck allocating memory to store connection state)
asciilifeform: that's not even to mention items like http servers where any joker can make you thrash disk and eat up 100s of msec per req
asciilifeform: whereas w/ pest station, a stranger can't even ping the box and distinguish ip from a dead one.
asciilifeform: given as it won't answer.
asciilifeform: nobody but a peer has any biz at all knowing that there's a box at $ip.
asciilifeform: likewise, you might be sending from $ip1 nao, but $ip2 on entirely diff. isp 10sec from nao, and the peers don't give a damn. they'll answer to the most recent one which spoke.
jonsykkel: ye this is cool property
asciilifeform: it aint an accidental property, jonsykkel , it's rather fundamental to the design.
jonsykkel: not very familiar with ddos or network stuf in general
jonsykkel: i see i see
asciilifeform: notion is to abstract away as much of the underling physical network as possible.
asciilifeform: cut cable? no prob., so long as your 2 other cables, or gsm modem, etc still live.
asciilifeform: and no need for complicated routers or whatever heavy gear.
asciilifeform: or say yer 'pesting' from a moving vehicle, and 'wardriving' wireless. also noprob.
asciilifeform must bbl
jonsykkel: very cool stuff indeed
jonsykkel: looking forward to shortwave pestnet
asciilifeform: jonsykkel: one could speculate re exotica, e.g. pirateradio boats anchored off coast and transiting pest traffic for btc-paying subscribers; but 'bridge too far' atm.
asciilifeform: ( potentially if yer rockefeller, could even Officially reserve spectrum for this. but again 'bridge too far', i doubt anyone will be doing this in near future )
jonsykkel: hehe
asciilifeform: the bigger picture point is that the 'single packet friend-or-foe' allows for very simple paying-members-only services, in principle.
jonsykkel: one day
asciilifeform: (i.e. stations where you gotta pony up to have yer key in its wot)
jonsykkel: ye will be very easy to interface to anything
asciilifeform: first gotta 1) 100% sane spec 2) 100% compliant implementation which can eat&shit at a reasonable line rate. (seems that these gotta be baked in parallel.)
asciilifeform: then and only then exotica.
jonsykkel: best way to do it anyway, impossible to plan too far ahed imo even for einsteins
jonsykkel: or quality of planning will suffer anyway
asciilifeform: 'Первым делом, первым делом - самолеты, Ну, а девушки? А девушки - потом.'(tm)(r) ('first, first business, are the airplanes, then, and only then, the chix'
jonsykkel: words to live by
jonsykkel: or smth
billymg: i've updated the pest info page at http://pest.bitdash.io to include a mirror of asciilifeform's latest 0xFC draft, as well as a mirror of thimbronion's prototype
billymg: accidentally ran the bot modified for pestnet in this chan, let's see if it's working correctly now
billymg: ok, back to normal
asciilifeform: billymg: neato, ty!
asciilifeform: billymg: loox a++ eye-friendly.
asciilifeform envisions 1 day possibly someone could make the whole protocol in wall poster form
asciilifeform: imho is mechanically simple enuff for this to be effective
asciilifeform: re: 0xfc -- considering notion that 'Text' perhaps oughta be renamed to 'Payload'; the orig. name may be misleading in light of binariola payloads being permitted in current ver.
asciilifeform: would ideally go along with a revision of the verbiage ( word 'payload' is in places used to describe entire Message with respect to a packet , in particular )
asciilifeform: this and other potential confusions oughta be polished away.
asciilifeform: 'Text' oughta refer strictly to the payloads of Messages carrying human-readable strings.
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-21#1067095 << cool idea, maybe i'll make some numbered prints and mail 'em out when the spec is finalized
dulapbot: Logged on 2021-11-21 16:42:38 asciilifeform: envisions 1 day possibly someone could make the whole protocol in wall poster form
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-21#1067097 << i was thinking the same re: the intro text, "Pest is a peer-to-peer network protocol intended for IRC-style chat."
dulapbot: Logged on 2021-11-21 16:47:03 asciilifeform: re: 0xfc -- considering notion that 'Text' perhaps oughta be renamed to 'Payload'; the orig. name may be misleading in light of binariola payloads being permitted in current ver.
billymg: perhaps, "...intended for decentralized packet routing" or something similar
thimbronion: billymg: love the landing page
billymg: thimbronion: thanks!
asciilifeform: billymg: right nao it's strictly for text chat; when we have 100% watertight function there, then can add some megalomania in spec, heh
asciilifeform personally favours wotronics (i.e. read/query ratings etc) as 'next pheature'
asciilifeform thinking that 1.2.7.4 oughta move to sect. 9.
asciilifeform: it disrupts the simple n00b intro of sect. 1 imho.
← 2021-11-20 | 2021-11-22 → ↑
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code