Protoocl Work Group Call #11 - near/dev/hub

Captions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.

1. 0:05

   hi everyone my name is l and I'm de contributor at near devhub I would like to welcome you to the 11th protocol workg group call where we'll be discussing and arriving at the decision on uh nap 9 that introduces stateless validation stage zero the agenda for today we will briefly go for the objective introductions uh do the disc discussions around 509 questions and answers and how to decision and how to get involved so uh to for for those of you

2. 0:37

   who have an attended such a call I'll quickly explain the process um there are four rows into this uh into this course and the process alter and champion who writes the proposals and suggests uh uh and new enhancement proposals moderator which facilitates the process which overall I'll be taking now the reviewer which is for smmes uh from the working group and the working group members who make the final decisions on on these

3. 1:09

   proposals so each uh each nap uh goes into through um three stages which is draft review voting which voting could result to being rejected or or approved you can learn more about those on the um on the GitHub link you can see below um in introductions the the the original author of the today's naap is Yun uh but of course they have been also other contributors to uh to this proposal and um other people have been working on it including voo Alexander and ton and Yan I'll be playing the

4. 1:51

   moderator role um the assigned reviewers are Michael and Marcelo and the approval the working group members are Marcelo Michael Martin and Bowen so um without further Ado um let the discussion of the nap 509 start and I would like to welcome Yun who is going to talk more about it you I um hope you can hear me well um I'll just talk about very high level overview um I mean obviously the nap itself was written by a lot of contributors especially from the technical aspects and that is why we

5. 2:34

   have a several member of the team here including a typhoon and Anton thanks for joining so basically Sal list validation is our way to achieve a phase two of charting that has been long overdue and uh the team has been working on this for close to a year at this point so basically at the high level uh what we are trying to achieve is uh the um no there is no node or validat then needs to track more than one shart uh that is the kind of like the Baseline segment we want to achieve and here with the Sal

6. 3:06

   this violation we are trying to do that um without relying on fishermen or wag or challenges um by um fulfilling the assumption that A Shard is secure if every single Chunk in that Shard is validated by randomly simpled subset of all validators so what it also means is now instead of like um each node each validators maintain their own state in local disk they'll get all the data they need to they need to validate incoming transactions over the network um and um

7. 3:42

   this uh St list validation also introduces several changes to the way things work on a near protocol Network including some role changes um so instead of having a chunk only validators and the uh block valid block producers SL validators we will have a new role um called the chunk validators which does not produce any blog word chunk um and also um there will be some Hardware requirement as well because of our um the new approach uh which

8. 4:13

   requires a uh state to be loaded into the memory so um we also share those details on like what kind of a new hardware requirements are needed uh for Dev validators and U the validators are aware of the necessary steps to do so um this is just a initial round of uh our approach meaning there are lots of optimizations and changes we plan to make after this initial launch so uh please be mindful that there are some known uh the uh pitfalls and the

9. 4:46

   security concerns that uh we plan to address in the upcoming release as well so I'll stop here great thank you um I like to um bring up uh one of the reviewers that are assigned for reviewing this uh proposal um I see that Marcel is not currently on the call Michael would you like to express uh your your opinion and um yeah how do you feel sure yeah happy to do that um so I guess first and foremost congratulations to the the near cor team for for all the work around

10. 5:26

    this like this is this is a huge change to the to the protocol but it's it's a welcome one we're always happy to see the scaling of of near be improved and as you mentioned this the the original concept with uh roll backs and challenges and fishermen was always com complicated to say the least and and Status validation is is a much more uh elegant solution to to this problem so I'm I'm super happy to to see

11. 5:57

    this great thank you and I I would like to invite B and W uh as head of protocol um we would like to see what are what is your opinion on the proposal uh and what are the benefits if you can talk a little bit more about the benefits and the concerns about this NE B yeah so uh I think a few major benefits right so one is that eliminates the uh need to uh Implement uh fraud proof and and uh chain back for uh phase two of sharding um and um that's

12. 6:34

    definitely a major benefit for um for the for the protocol um and another major one is that uh um yeah the uh the state now is entirely well pretty much for practical purposes entirely memory uh whether it's on the trunk producer side or it's on the trunk bodor side so uh when essentially that means when a chunk is applied uh no matter on like which type of validator node uh it's always uh State access always in memory and this is going to uh dramatically improve the performance uh

13. 7:14

    of uh of the network um and uh as we know the state read and writes uh is the biggest performance bottle neck that we have today uh so very happy to see that being addressed as part of the uh the same proposal and then beyond that there is also benefit on decentralization of the network because the validators are now much uh supposedly much more light lightweight because they don't need to actually uh store any State uh locally so they can afford uh to wrong with less

14. 7:46

    expensive hardware and um also they will be validating one Shard at a time uh versus uh today um so yeah that's uh as Michael pointed out it's a major uh change to the protocol and uh uh supposedly will take uh the scalability of near to the next level uh now regarding the concerns I think the it's mostly that um this is uh the first uh iteration and there are still uh many things that could be improved it's not necessarily that there

15. 8:20

    there are something inherently wrong with the current implementation I just said for example the state witness size is not necessarily optimized and there may be certain uh overhead uh regarding the communication uh of witness to uh to validators that are uh not optimal and because of that uh it doesn't uh fully um I would say doesn't fully realize the potential of the new design um but that would be um work for the for the future

16. 8:51

    and um um and I think the proposal as is uh is a very good Milestone uh for for the entire uh near protocol do any any of the other working group members want to express any uh benefits or concerns with the current uh proposal Martin so first I will second what Michael said that this is huge I mean this is probably one of the biggest changes to protocol comparing to the all the other naps that we ever discussed so

17. 9:33

    yes it's a great opportunity at the same time uh there is this tingling sensation that like holy there are like so many small things that we hopefully covered if you if you go through this snap it's also for you Vincent if you go to the snap you can see very detailed like analysis of pros and cons and statistics etc etc so most of the discussion kind of happens there what we are here in this thing is like the last year you know official signing of this rather than the

18. 10:03

    whole discussion so once again amazing job Yun and the team and yes uh let's get this thing going all right amazing um as per the protocol for um approving those uh proceeding with those NPS uh the work group members are voting on um are voting on approval or rejecting and uh we need a majority of the votes to go either way and the working group members have uh U discussed uh this on the public channels including on the net itself on that can be found on GitHub

19. 10:43

    and uh we have arrived at decisions to uh approve this NP so congratulations on that everyone amazing work and thank you very much for that uh I would like to um to get uh Bowen if you have like any final words or like does do you change your opinion or have something to add finally um because the majority of U of the members have voted for uh yeah I I yeah I wholeheartedly support this uh nap and as Marson said yeah it's a a huge effort and uh yeah

20. 11:22

    great work to uh the entire team and um yeah it's great to see it happen amazing all right um I so congratulations to everyone and thank you for the amazing work uh I would like to continue with questions and answer but before getting into it um I would like to get to know what would be the next uh steps for and like what is what more work needs to be done in order for this n to be um

21. 11:54

    applied Yun if you can take this one um we have addressed all the Baseline uh the implementations we that was discussed and mentioned in the nap so right now the sales validation uh the 2.0 release is released was released to testnet um and has been under testing uh with the various uh different network we're running with the community as well as internal testing and our Target main net release date is as of now August 5th

22. 12:25

    so we have about two more weeks until we get there uh and um of course this state may move around depending on what we find in test Nets and other test but um it'll just continue going and um like you know like we bunch of improvement and optimization plan are part of our road map in age2 of 2024 uh so I think we are in a relatively good spot as of now amazing so just to clarify that um

23. 12:55

    the proposed changes made uh if everything goes to plan may take place into the late 2024 correct um the proposed plan is uh to take a place in mainnet in August oh amazing yeah so in two weeks yes amazing um any other questions uh for from the community or if uh any of the people attending this would like to ask the author or the work group members I I have a question but is about the process not this particular

24. 13:39

    NP all right I discussed earlier we can do this async and I will link you to some resources or how to um how to get started uh on the process well the question I have ex actually very specific um I have a net proposal and I still awaiting to get reviewers assigned so I'm not aware if there's anyone particular with the responsibility to assign them or I have to find people who are interested in doing the review uh I I wonder how to

25. 14:12

    best you know move forward with that yeah the moderators should review and this and sign reviewers right uh Ora I don't know like who what y what is going on there yeah so we've uh recently like made some transitions on the moderator team so there's probably like we're probably delayed on some NPS um can you drop the link to the NP you're referring to um and I'll make sure that the team

26. 14:45

    takes a look at it uh we do a we like a weekly backlog review basically to go through all the nips and kind of figure out if one they meet the actual requirements and if not we give feedback so if you haven't received that apologies like I know that there's a little bit of a backlog but we'll we'll make sure to do that and then if it does meet the requirements then we would reach out to the appropriate work group and get them to do the reviews so so

27. 15:08

    yeah I will make sure that our team kind of takes a look at that and uh we also have some other NPS that we need to review uh as well so we're going to get back on track uh in the next week I I'm going to put it in the chat perfect thank you actually this is this is something I I discuss in the Twitter group uh so I I asked you let's say I asked permission before you know before doing the pulling the pool request I didn't just throw it there let's

28. 15:40

    see amazing okay yeah it looks like oh I see okay yeah yeah we'll take a look at this it looks like the the there wasn't a specific work group that was tagged it was just like a general work group and so nobody reive the notification for it so uh we'll make sure to see uh yeah which which group we need and whether it makes sense for this protocol work group or another one to take a look so it's quite a a unique one so I I I is

29. 16:13

    I it's no wonder that it wasn't easily assigned to anyone all right um well uh with that I would like to thank uh everyone for uh attending today's call uh for the author of altering this NP subject matter expert work group members in the community um especially the community thanks uh for vent for your work and dropping by it's always awesome to see people getting involved and

30. 16:43

    suggesting uh proposals to make our like life quality life better uh so looking forward to three your proposal as well if you want to get involved check out the devhub website check out for the different communities that that we have if you want to get to know more about the Naps you can check the GitHub na repo under the near organization and if you want to see on the upcoming naps just follow the link

31. 17:20

    on screen and yeah if you want to continue Asing discussion uh discussions and get more information about uh this NS and other protocol related things join our telegram Channel and see you there thank you very much