-
AMeMario
What is Tari?
-
AMeMario
Monero inflation bug? boards.4channel.org/biz/thread/26234830
-
sgewehr
What kind of person steals from own community?shorturl.at/jmDM4Your own leaders are laughing at how stupid you are for falling for thier 'Magical Crypto Friendship'.
-
sgp_
Meeting in ~15
-
sgp_
MRL meeting time
-
sgp_
-
sgp_
1. Greetings
-
ArticMine
Hi
-
TheCharlatan
Hey
-
sgp_
ping knaccc sarang dEBRUYNE Isthmus moneromooo
-
sgp_
some people may show up after Biden's speech (happening now)
-
zkao
hello
-
sgp_
before then, I will give a brief BP+ update
-
sgp_
sarang gave me some emails to contact asking for SoWs, and I will email them today
-
lederstrumpf
hi
-
sgp_
the other BP+ audit was moved, funded, and started since last meeting so that's quick
-
sgp_
any questions on BP+ audits before we move on?
-
ArticMine
Any idea on timeline
-
sgp_
"about 1 month"
-
ArticMine
Thanks
-
sgp_
okay, let's actually skip to v15 ringsize if sarang isn't here to discuss triptych
-
ArticMine
Yes I agree
-
sgp_
-
sgp_
there are a few different opinions on this one
-
sgp_
I personally think a bump from 11->15 is most appropriate
-
sgp_
ArticMine: you suggest a higher number ~19-25 still, right?
-
» dEBRUYNE here
-
ArticMine
Yes 21
-
sgp_
I took a look at the CPU benchmark numbers you shared
-
sgp_
the increase you noted was for the high-end hardware that's often out of stock at the moment, from what I can tell
-
sgp_
looking at benchmark per dollar, my best estimate is 25% increase
-
dEBRUYNE
Wouldn't 21 raise the verification time significantly? Presuming we merely implement Bulletproofs+
-
sgp_
I don't have specific numbers for 21
-
sgp_
but yes
-
ArticMine
I saw your response. I find the high end as the best indicator. Basically "value" is more often than not junk going back to the pc jr in the 1980's.
-
ArticMine
One is better off with slightly older high end from a cost vs return perspective than the newest "value"
-
sgp_
15 would increase verification (of the signature and balance, NOT the entire transaction) by 35%
-
ArticMine
It is a 2x increase for the signature verification
-
sgp_
17 would increase 54$
-
sgp_
*54%
-
ArticMine
for 22
-
ArticMine
but also it bring the size close to where Triptych would come in
-
ArticMine
One thing to keep in mind is that one we go into the dynamic blocksize we have to be very careful how we increase the tx size
-
ArticMine
It can be done, but it can require modification of the long term median right after the fork
-
dEBRUYNE
I'd prefer to impement a ring size bump with Triptych
-
dEBRUYNE
implement*
-
sgp_
I think going for a 100% increase in verification is optimistic
-
sgp_
I fear it's too aggressive
-
sgp_
especially since it doesn't specifically address any threat model
-
ArticMine
I am figuring an on triptych coming in with a ring size of ~64 at least
-
sgp_
yeah, 64 or 128 (I'll probably push for the latter thinking ahead)
-
ArticMine
So a more gradual increase does make sense
-
sgp_
well, there are pretty major costs right now for minor benefits
-
sgp_
knaccc is for 16 I believe for the churning numbers he shared
-
ArticMine
By the way wownero is running ring 22
-
ArticMine
Yes that is based upon an arbritary mix set of 1,000,000
-
ArticMine
after 4 churns
-
sgp_
yeah I agree it's quite arbitrary
-
ArticMine
SO there is a gain with each level of increase
-
sgp_
just to stress again why I am proposing an increase to begin with:
-
dEBRUYNE
I think at 64 the verification time would remain approximately similar right?
-
sgp_
11 is still "safe" in my opinion, but the buffer has narrowed more than I feel comfortable with personally
-
sgp_
dEBRUYNE: approximately, yes
-
sgp_
-
sgp_
sorry that's size not verification time
-
ArticMine
The crossover is well above 15 closer to 21
-
ArticMine
between CLSAG and Triptych
-
ArticMine
more if we extrapolate Triptych to 64 or 128
-
ArticMine
Verification time is linear with anonymity size
-
sgp_
dEBRUYNE: do you still think keeping it at 11 is best?
-
ArticMine
It is still linear with triptych
-
dEBRUYNE
As far as I can see, the probability of Triptych getting implemented is fairly reasonable
-
sgp_
yeah but is the slope less? iirc the slope is less
-
dEBRUYNE
So I am wondering if this discussion is somewhat fruitless in the first place
-
sgp_
well, everyone else isn't confident that triptych will happen very soon
-
sgp_
signs point to v15 without triptych
-
dEBRUYNE
No, but I think end of the year is still a reasonable target
-
dEBRUYNE
And to fork merely for Bulletproofs+ seems kind of wasteful
-
ArticMine
There are issue with multi sig wallets
-
sgp_
so in any case we still have a time v15 without triptych to select a ringsize
-
dEBRUYNE
ArticMine: Yes, but nothing that cannot be overcome as far as I can see
-
sgp_
it can be overcome but just not hastily
-
sgp_
it's a huge change
-
ArticMine
^ that is my point
-
dEBRUYNE
Sure, but why the rush to implement Bulletproofs+
-
ArticMine
It take time
-
sgp_
BP+ is a quite minor change by comparison
-
dEBRUYNE
The optimizations are marginal
-
sgp_
take the efficiency gain today
-
ArticMine
They do add up
-
sgp_
taking the gain doesn't delay triptych
-
dEBRUYNE
We also have to remind ourselves that the ecosystem grew a lot
-
sgp_
if we were pushing triptych back, I would agree
-
dEBRUYNE
Not sure if they change the transaction format, but in that case wallets would have to add logic for Bulletproofs+ too
-
sgp_
but that doesn't appear to be the case
-
dEBRUYNE
(the ones that use custom code)
-
dEBRUYNE
Also if we add a HF in say August, we cannot do one for Triptych in say November/December
-
sgp_
triptych frankly will not happen in 2021
-
ArticMine
There is also issue 70 in mrl
-
ArticMine
Tha twill need a hard fork
-
dEBRUYNE
That seems like a bit of a preliminary inference
-
sgp_
August is also later than it needs to be for BP+
-
sgp_
May maybe?
-
dEBRUYNE
5-6 months after our previous HF?
-
dEBRUYNE
Which didn't go particularly smooth either
-
sgp_
7 months I guess, could push back more, not technical reason to however
-
sgp_
*no technical
-
dEBRUYNE
Apart from the fact that it puts a strain on the ecosystem
-
dEBRUYNE
It's what was discussed in the blog
-
dEBRUYNE
-
sgp_
I simply do not see triptych as being implemented as optimistically as you
-
dEBRUYNE
Perhaps sarang or moneromooo can weigh in, but I think having 9 months is definitely doable
-
sgp_
well, moo said they don't want to touch that code, so we would need someone like sarang to start running with it
-
sgp_
then we would need to notify everyone of the upcoming change
-
sgp_
then we would need to make super clear resources for users of multisig wallets
-
selsta
BP+ will require changes to HW wallets, right?
-
sgp_
and give them time to coordinate and "convert"
-
dEBRUYNE
As far as I know they don't necessarily need to convert existing funds
-
dEBRUYNE
Merely generate a new wallet for Triptych funds
-
sgp_
we discussed a major issue last week, where they need to "convert" to a non-multisig wallet to spend triptych funds
-
sgp_
that conversion period will take a lot of education
-
sgp_
point is, unless we have someone ready to run with the idea, we're kinda hoping for something that isn't even penciled in currently
-
sgp_
we can work on getting everything outlined
-
sgp_
but we don't currently have anyone working on this
-
sgp_
nor has sarang said "I will do this if I get the $$$"
-
ArticMine
BP+ does not need any of the conversions that Triptych will need
-
ArticMine
So I way we need to proceed with BP+
-
dEBRUYNE
I suppose we can assess the viability again in a few months, but it also seems to be a bit preliminary to say that we can fork BP+ in summer
-
sgp_
well, at that point we lose the efficiency gain anyway haha
-
dEBRUYNE
Not sure I am following
-
dEBRUYNE
The gain starts counting from the moment it is implemented
-
sgp_
I don't want BP+ to be something we delay and hope that triptych will come together in a few months
-
sgp_
if we want to commit to triptych, it will take significant active effort
-
sgp_
else we're delaying bp+ for no reason
-
sgp_
*no significant reason
-
dEBRUYNE
I meant that in 1-2 months we probably have a fairer assessment of the viability of Triptych in 2021 (end of)
-
sgp_
okay, so what are the specific next steps we need to take to figure that out
-
ArticMine
... but even with say Triptych at the end of 2021 it still makes sens to go ahead with BP+
-
ArticMine
sense
-
sgp_
ArticMine: if it actually is 2021, I'm less for BP+ right now
-
ArticMine
It depends when in 2021
-
u29601mg6ba93j[m
I think a definite gain soon (BP+ and ring size increase) is a great hedge against the real possibility that triptych will take longer than we hope to implement. We can argue about which CPUs to benchmark but ArticMine is right that they have advanced a lot over the past 2 years. That should be accounted for when looking at verification times. The tx size increases are not large enough to be problematic
-
u29601mg6ba93j[m
for the ring sizes we are discussing
-
sgp_
to dE's point, a 5% gain isn't worth an upgrade headache for that time period, at least imo
-
sgp_
that being said, if we feel the ringsize needs to be increased before EOY, then our hands are tied and we should do one earlier
-
sgp_
quick vote
-
ArticMine
I see as there are enough incremental inprovements to make it worthwhile overall
-
sgp_
Do you believe it's critical to increase the ringsize before the end of the year? Y/N
-
ArticMine
Y
-
u29601mg6ba93j[m
Y
-
sgp_
N
-
moneromooo
triptych will depend on whether sarang or someone with a similar level of clue either adds it to monero, or adds a *production quality* python implementation that I can convert to C++.
-
sgp_
ArticMine u29601mg6ba93j[m: why do you feel that an "immediate" (very soon timeline) ringsize increase is necessary?
-
ArticMine
I am also concerned about the regulatory issues
-
sgp_
is that the main reason or a side reason?
-
ArticMine
It is a major reason. There is also the multiple attack vectors mining, CTRs, etc
-
ArticMine
They can have a cumulative effect
-
sgp_
I personally don't see a regulatory difference 11 vs (eg) 25
-
sethsimmons
<sgp_ "Do you believe it's critical to "> Y
-
sethsimmons
Sorry I'm late 🙂
-
gingeropolous
Y for me as well
-
sgp_
sethsimmons gingeropolous why do you feel it's essential to increase the ringsize Q2?
-
u29601mg6ba93j[m
Mostly because we are all guessing about how soon Triptych can be implemented. Moneromooo just mentioned one key dependency (availability of Sarang) that we cannot be certain of yet. We cant go back in time. Skipping BP+ and ring size increases now, will become a regret if Triptych gets delayed for 1 year or more
-
u29601mg6ba93j[m
* @sgp_ Mostly because we are all guessing about how soon Triptych can be implemented. Moneromooo just mentioned one key dependency (availability of Sarang) that we cannot be certain of yet. We cant go back in time. Skipping BP+ and ring size increases now, will become a regret if Triptych gets delayed for 1 year or more
-
sethsimmons
<sgp_ "sethsimmons gingeropolous why do"> EOY is still a long ways off and there will surely be new threats to the network by then. As you’ve said, I feel that 11 is OK for the moment most likely, but I definitely would like to see a bump before EOY
-
sgp_
is there a date everyone has in mind for "11 really needs to go by then"
-
ArticMine
ASAP
-
gingeropolous
1. triptych implementation timeline unknowns. 2. using every opportunity available to strengthen the network within reasonable technical limitations. This goes to my viewpoint that we have enjoyed seemless protocol upgrades and their is no sign of ossification, but ...
-
sethsimmons
Takes a good bit of time to plan and deploy a HF across the ecosystem, so ASAP for me as well.
-
sgp_
last chance to respond on this q, then I'll ask another
-
binaryFate
lurking, I support ASAP for >11 too
-
UkoeHB__
I do not think multisig wallets would need to convert to non-multisig to spend Triptych funds, unless Triptych-friendly multisig isn't implemented (it's a non-trivial task)
-
sgp_
UkoeHB__: can you read sarang's statements form last meeting and clarify?
monero-project/meta #542
-
sgp_
second q time
-
Lyza
dunno how relevant this would be considered, but there's been a lot of patches recently to address ongoing attacks and a consensus change would make them effectively mandatory which could be a plus
-
sgp_
if the decision is to increase ringsize ASAP, how early can we have v15
-
sethsimmons
As soon as BP+ is audited and all issues are corrected. Maybe mid-late April?
-
selsta
earliest July IMO
-
sethsimmons
That may be too soon though.
-
dEBRUYNE
As a general comment, I don't like this push for a hard fork 'ASAP'
-
dEBRUYNE
We informally decided to slow them down due to the strain on the ecosystem and now I see a few people here pushing again for one within 5-6 months of the previous one
-
sgp_
is anyone for a sprint? thinking like April. Just trying to feel this out
-
dEBRUYNE
Hard pass from me
-
sethsimmons
ASAP as in "as soon as we can reasonably deploy one across the ecosystem".
-
ArticMine
Yes of course
-
UkoeHB__
sgp_: afaict he said two things A) Triptych multisig is non-trivial and may not work on hardware devices, B) he thought maybe it reveals a private key to all participants but appeared to walk that back here
monerologs.net/monero-research-lab/20210115
-
ArticMine
My take ~9 months since the last HF
-
selsta
that would be July / August
-
ArticMine
Yes that makes sense to me
-
knaccc
btw I don't have a strong opinion yet on ring sizes, because I'm not sure what the stats will look like on verificaiton times for a year's worth of blockchain
-
sethsimmons
I'd get behind July/August.
-
sgp_
I agree that if we are waiting until that time, we have a few weeks to put together a triptych plan
-
sethsimmons
Don't want to make things too painful for ecosystem participants.
-
gingeropolous
so July / august would put us at April 2022 for triptych if we're gong for ~9 months in between
-
sgp_
If we can get triptych in Dec 2021 actually, then I personally think that's better than BP+ in July/August
-
dEBRUYNE
Does BP+ change the transaction format by the way?
-
dEBRUYNE
UkoeHB__ ^
-
sethsimmons
<sgp_ "If we can get triptych in Dec 20"> If that's possible I could get behind that. But that seems highly optimistic.
-
sgp_
the main reason to do v15 quickly is if the current ringsize is determined to be unsafe (think like an emergency)
-
selsta
dEBRUYNE: yes, afaik Trezor will have to update their firmware and Ledger update their app
-
UkoeHB__
not sure dEBRUYNE
-
selsta
at least I think so
-
sgp_
if we're thinking an increase many months away, that doesn't seem to me like people think it's an emergency
-
gingeropolous
but BP+ also works for triptych right?
-
sgp_
gingeropolous: yes
-
u29601mg6ba93j[m
<selsta "that would be July / August"> 9 months seems like a good compromise. I would prefer faster but balance that preference to what dEBRUYNE said about challenges for ecosystem and our goal to decrease fork frequency
-
sethsimmons
<sgp_ "if we're thinking an increase ma"> Its not an "emergency", no.
-
sethsimmons
Necessity if Triptych is not possible this year, yes IMO.
-
sethsimmons
But doesn't require wrecking the ecosystem to get it quickly.
-
Lyza
echoing knaccc 's concern.
-
Lyza
it keeps being said that the given verification time increases are not for the entire transaction, but those are exactly the numbers that seem most relevant to the verification time discussion
-
dEBRUYNE
<sgp_> the main reason to do v15 quickly is if the current ringsize is determined to be unsafe (think like an emergency) <= Is it?
-
TheCharlatan
Since bulletproofs+ does not only change proof verification, but also proof creation, it will require a wallet side update.
-
dEBRUYNE
I am skeptical of that statement to be honest
-
sgp_
if I can summarize things a bit just so we don't talk in circles
-
dEBRUYNE
TheCharlatan: Thanks. That means MyMonero etc. need to upgrade too, which will be a hassle
-
sgp_
we are a few months away from the soonest anyone wants to have a network upgrade
-
dEBRUYNE
Especially if they have to upgrade a few months later again for Triptych
-
sgp_
that gives us a small amount of time (not a huge amount!) to feel out a triptych action plan
-
sgp_
if within a month we don't have a triptych plan, seems like most people are for BP+ and a ringsize bump in July
-
sgp_
since people don't want to drag their feet forever
-
sgp_
if however triptych is actually doable by EOY, I see strong arguments for pushing v15 to Dec to not delay Triptych by 6+ months
-
Lyza
agree
-
sgp_
is the above in contrast to anyone's thinking?
-
gingeropolous
that logic seems sound to me as well
-
sethsimmons
No, agreed here.
-
UkoeHB__
it's reasonable
-
sethsimmons
Is initiating HF planning in late Feb/early march sufficient time for forking in July?
-
sgp_
imo yes, we need 3 months on the planning side for a relatively "small" upgrade like BP+
-
ArticMine
That is reasonable. Aim for July / August and if Triptych is viable by December then delay to include Triptych
-
gingeropolous
i would put forth that if triptych is not seen doable by EOY, we consider a ringsize bump without BP+. This would put the least strain on the ecosystem (no 3rd party upgrades etc) but still get everyone on the same software, which re: p2p attacks, is good
-
sethsimmons
OK, good with me then.
-
sgp_
okay, I will make a visual showing the rough timeline options then
-
sethsimmons
<gingeropolous "i would put forth that if tripty"> That's a good alternative but I'd prefer to make the fork "worth it" with some important upgrade like BP+
-
sethsimmons
It causes ecosystem strain either way simply HFing, although without BP+ would be lessened.
-
gingeropolous
but BP+ is gonna cause 3rd parties to have to fiddle, and then they're gonna have to fiddle again with triptych
-
ArticMine
There is also MRL 70 fee stabilization
-
gingeropolous
i guess good info would be how ready are 3rd party things to fiddle
-
sethsimmons
<gingeropolous "but BP+ is gonna cause 3rd parti"> How significant is this fiddling?
-
sgp_
any immediately pressing matters to address at this meeting?
-
gingeropolous
ledger, trezor, mymonero
-
selsta
all mobile wallets too
-
sethsimmons
They all have to update for HF anyways.
-
u29601mg6ba93j[m
<ArticMine "There is also MRL 70 fee stabili"> Do you feel that should be done by July fork or can it wait for Triptych?
-
TheCharlatan
gingeropolous I don't think that is true r.e. no 3rd party upgrades if only the ring size is bumped. The wallets still need to update to create larger transactions.
-
sgp_
to reiterate, my personal action items are: 1) contact for second BP+ audit, 2) make visual of timeline options
-
ArticMine
It should be done by the July fork
-
u29601mg6ba93j[m
> <@freenode_ArticMine:matrix.org> There is also MRL 70 fee stabilization
-
u29601mg6ba93j[m
* Do you feel that should be done by July fork (if applicable) or can it wait for Triptych (perhaps not until 2022)?
-
TheCharlatan
I don't think bulletproofs+ will bring much more fiddling once it is reviewed and in a stable state.
-
gingeropolous
ok
-
ArticMine
because we could go into the adaptive blocksize and then the recent attack could have bite
-
TheCharlatan
The only thing that significantly changes for the hardware wallets will be some message sizes. The bulletproofs are anyway already done by the monero wallet, not the device firmware.
-
u29601mg6ba93j[m
<ArticMine "because we could go into the ada"> That is a great point, especially if adoption increases significantly this year
-
selsta
TheCharlatan: so less work than CLSAG?
-
selsta
that would be good
-
TheCharlatan
Yes, definitely less work than CLSAG.
-
ArticMine
basically if we are into the adaptive blocksize an attack for a matter of hours can trigger a fee increase that can take months or even years to correct
-
ArticMine
There is a real incentive here for an attack similar to what happen recently
-
ArticMine
but it only works if we are in the adaptive blocksize not in the penalty free zone
-
UkoeHB__
ArticMine: have you published your latest recommendations yet?
-
ArticMine
I have them finalized not published yet
-
UkoeHB__
cool I look forward to it
-
ArticMine
I expect within a week
-
u29601mg6ba93j[m
Since it is coming within a week. Can we have another vote next week(same as before except adding MRL70 to the July Bp+ and ringsize increase fork)?
-
ArticMine
We can meet in a week
-
ArticMine
same time
-
dEBRUYNE
Sounds good
-
sethsimmons
Thanks all!
-
ArticMine
Thanks
-
gingeropolous
Thanks everyone!
-
sgp_
thanks!
-
gingeropolous
ArticMine, is your current work based on testing the existing software? I'm thinking of putting in the effort to actually test the adaptive blocksize stuff using release software
-
ArticMine
It is about issue 70 in MRL.
-
ArticMine
-
ArticMine
Now combine this with the recent attacks
-
ArticMine
It only works f we are naturally deep in the adaptive blocksize
-
ArticMine
if