…
…
Are these NP among the owners who have never registered as node providers and own 63 or so of the unassigned nodes that are receiving remuneration? In my opinion, they should not be getting remuneration beyond the initial 4 year agreement if they cannot be contacted and are not willing to follow the procedures required by all node providers that have been established for the next 2 years. That seems like reasonable enough justification for removing them. Is that what is happening here?
Yes, precisely. I would not be comfortable having for 2 additional years NPs that I can’t communicate with. Or that do not have sufficient technical skills and need to rely on 3rd party for any kind of technical work. I would personally be very happy if someone from the community would take over (buy) these nodes and continue to operate them. Ideally in a new geographical area and legal jurisdiction.
Hello @timk11
We are currently offboarding four DCs that were a unique situation where DFINITY owned the servers, but other NPs were running the operations of the nodes. We are not continuing that model, so these four DCs are being offloaded. Three of the DCs worth are being converted for our internal use, and the fourth DCs servers were part of our auction.
This thread disuccesses the voted-in terms under which some Gen-1 NPs (including ourself) will be reducing and offboarding some of their nodes. To my knowledge, all of them have found private buyers for those nodes (as we did) who will bring those nodes back up on the IC, so that the IC node count does not shrink by any more than the 3 DCs mentioned above.
Remember… Gen-1 NPs signed up and began their DC contracts before Genesis, as the IC had to be set up and tested on these nodes in order to launch, and it took time. Their 48 months started when the DC building began on their DCs, so there was always a delay between that start point and when the nodes in that DC actually were up and running and visible in the dashboard.
The NPs in the situation mentioned above and two more larger NPs are reaching the end of their 48 months this month. They were the first DCs that we used for starting the IC (after our own Zurich DC). All the rest of Gen-1 NPs will reach their 48 months at the end of January. According to the terms described in this thread (and the wiki here). Those list the steps that they and the buyer will have to follow, along with the proposals they will have to make. So each of the NPs who has more than 42 nodes will make a post in that forum thread announcing which DC they are keeping and which they are selling to another buyer and who that buyer is, so there will be a handful more of these sets of proposals by the end of the year.
I hope this answers some of your questions!
There are not any NPs who cannot be contacted. Almost all have joined the Element channels, and the few that haven’t are in frequent communication with me about the situation. You will also see in the forum thread and wiki listed in my message above that posting their element thread is a requirement to run their nodes beyond their 48 months. No worries!
Thanks for the info @katiep and @sat, and for asking really great questions @timk11, @wpb and @ZackDS.
This comment made me want to ask for evidence that the NPs have consented to these proposals. In general I think the burden of proof should be on the proposer to justify the proposal in a way that can be verified (and to highlight/provide a means of that verification).
Is this a fair perspective to have?
I’m not clear on what you’re asking.
The proposal that I linked to about rewards post-48-months was put forth by some NPs themselves. It was voted in by the community. That is how the IC works, and all NPs understand that when they become an NP… that any voted-in proposal can change things.
Please clarify if that does not answer your question.
I’m asking if proposals should be raised that require the community to scout around for evidence that the proposal is verifiably justified. By comparison,
Sure, and the voting community is what makes this practical. But this only works properly if the votes are informed, and the proposals are verifiable. I’ve not had a chance to dig into this properly yet. Taking a random proposal…
Where can I verify that the SJ1 DC specifically is happy to be removed from the network on the basis described?
I see your point, and you’re right, that isn’t posted publicly anywhere because these were being operated under an agreement that was made prior to Genesis when there was not a community to consult.
DFINITY is the one who owns these servers. The NPs for these four DCs (AT1, TP1, SJ1, FM1) were operating them under a private agreement made prior to Genesis for 48 months, which ends Nov 30. These NPs have been completely responsible for running them, (DFINITY has not had any access and does not hold the DC contracts.) Their use of the servers ends Saturday, and we are therefore reassuming posession of the servers that we own.
Thus, these removal proposals are more in the nature of a cleanup than anything else. The servers are getting shut down by the operators and us regardless, according to the terms of our private agreement, but it will be cleaner if the subnets are removed prior.
ETA on Dec 13 so this update can be easily referenced and found in continuting proposals:
Thanks Katie. That’s exactly the information I was looking for. There are 5 DCs covered by this set of proposals (at1, dl1, fm1, sj1, tp1), so to check my understanding - Dfinity owns all the node machines in these 5 DCs, NPs operated them under an agreement, and documentation of the agreement isn’t available but it was established that the arrangement would last for 4 years and end on Nov 30 (this Saturday). Is that all correct?
That is correct for the nodes in AT1, FM1, SJ1, and TP1.
That is not correct for DL1. That one is owned by the NP who posted in the forum here. According to that post, he is keeping DL1 and LV1, which means he’s selling the nodes in PL1 (the dashboard shows that he owns all three.)
Note that there will be a bunch of proposals over the next several months relating to all of the Gen-1 DCs, but the NPs themselves will be submitting the majority of them. If you are interested in a summary, I provided it in this thread.
@sat Can you explain why the node is being moved from DL1?
@katiep I was just following your earlier instructions. Did something change in the meantime?
Looks like I misread DL1 instead of PL1, so I guess we need to switch. I’m sorry!
FYI, this is a DM. It’s hard to tell because DM titles that are initiated from a comment in a public discussion get the same title as the forum thread by default. I just renamed it so it would be easier to identify. My apologies for the confusion.
Here is the DM that I sent to Katie before her response referenced above. I’m sure she will post her response, after which I will also post my additional response…
Hi Katie. Would you please clarify a detail here? Are there any nodes that are currently receiving remuneration rewards that will no longer be receiving remuneration rewards within the next several months? I have this question in general for all node providers.
The fact that some node providers were operating DFINITY nodes and DFINITY is reassuming possession of the servers makes me wonder if these nodes that are currently receiving remuneration are going to be leaving the network as remuneration recipients.
If there are any nodes that will no longer be receiving rewards for any reason (e.g. DFINITY getting out of the node provider business, node providers leaving the IC, etc), then I would like to know if I can start a discussion with the community and submit a Motion proposal to the NNS to make the argument that there is a higher purpose for keeping those nodes slots in service than there is for removing them in the name of decreasing inflation. Specifically, the only reason anyone is performing technical proposal reviews is because DFINITY has been willing to provide grants. The work we are doing helps advance decentralization, but the grants offered by DFINITY are not enough. There are many people who applied for the grants who are left on the sideline because of this lack of funding. Node provider remuneration could be used to solve this problem.
We are about to see periodic confirmation implemented as well as governance canister and NNS dApp changes that will make it easier to find and choose people and organizations who are performing the work of proposal reviews, yet there is a paltry number of people and organizations that have had access to the incentives necessary to put in the work and develop credibility as a valid Followee option. I would like an opportunity to make every node position that is being given up available to help fill this gap by harnessing the remuneration to pay people for technical proposal reviews. I can do this through CodeGov. In fact, it’s the only contribution I can make to governance since I’m not a developer who can perform the reviews. However, I can recruit developers and I can administer grants. It doesn’t have to be just CodeGov.
Even if there is a time gap before new servers can be acquired and onboarded, there is still tremendous value in considering this proposition. I think there is a chance it would be well received by the NNS, especially if DFINITY is supportive. It could be community driven. It would take time to order and build these node machines. I know people who are willing to help spec the hardware, provide racking instructions, and travel to perform the networking and node setup on site. As you know, I don’t have personal funds to purchase servers. Hence, the only way to obtain servers is to raise capital. That would have to be a public effort where community members are given the opportunity to contribute to the fundraising in return for attractive returns from the remuneration. It would be a golden opportunity to find out just how important decentralization is to the ICP community. The number of nodes onboarded could be proportional to the capital raised. The funds could be used to support people and organizations that are currently active or willing to become active in roles that advance decentralization of the internet computer…technical proposal reviews, critical infrastructure development, decentralization focused tooling, node provider support, etc.
I don’t know how many nodes are being given up, but I would like to include all of them in the discussion even if there are more than 42. Of course, I would like CodeGov to be able to take over as many as possible, but it would make a lot of sense for them to go to many people and organizations. I can still coordinate the public discussion and whatever fundraising is required even if the recipients of the node slots are not just CodeGov. In fact, I could work to negotiate the data center contracts in many different geographies, but scope it such that many different node providers are slotted to the same racks in those data centers instead of it being just one node provider with many nodes in the same rack in the same data center. That way we take advantage of scale while still improving decentralization.
I wouldn’t want to start this kind of discussion unless I know it is supported by DFINITY. I’ve been around long enough to know that its a waste of time and effort pursuing ideas that are not supported by DFINITY. I think this kind of community based decentralization is what DFINITY wants, but it can’t happen without support. Hence before starting a discussion, I’d like to know if an idea like this has a chance of being supported by DFINITY.
By the way, I know I’ve been taking up a lot of your time lately on these discussions. I apologize. I just see node provider remuneration as a solution to so many problems that we face at CodeGov that are caused by a lack of sustainable funding and a dependency on grants from DFINITY. It a problem that I feel first hand as the founder of CodeGov, but I think we are just one example of a bigger problem with the path to decentralization of the IC. I’d really like an opportunity to become truly independent and to bring as many developers along as possible (CodeGov or not doesn’t matter to me). Hence, I don’t want to leave any stone unturned. I appreciate your willingness to consider my proposals.
My request also applies to these 63 unassigned nodes in the event any of them are not planning to continue as node providers and have not sold their node slots to someone else. However, I realize that you may have already addressed these 63 nodes with this comment below. Perhaps Bjoern should remove this section of his proposal if there are no nodes that need to be removed due to not being registered. Instead of being removed, I think this idea should be considered of allowing the NNS to decide if these node slots can be adopted by others who are performing work for the protocol if they can come up with new servers.
@Lorimer @sat @SvenF @paulm @bjoernek @ld-dfn1 @bjoern @Luis @mompo @lara
Hello Wenzel!
Are there any nodes that are currently receiving remuneration rewards that will no longer be receiving remuneration rewards within the next several months?
Yes indeed. There will be a significant decrease in rewards being minted over the next few months. All of this will be happening under the terms of this same thread, but this is a high-level breakdown for what you’ll see happening as far as rewards go and proposals:
All of the above should result in a roughly 33% reduction in rewards being minted to Gen-1 NPs, over the next four months.
Regarding the rest of your proposals, there isn’t room to do that with existing servers, as DIFNITY needs to repurpose the rest of these servers, and repupose them in the United States where they already are (and where the IC does not need more nodes). We already have people with travel plans in place to move these servers next week. I hope you understand that it does not make financial sense for us to sell servers that we have and buy new, when we can simply repurpose what we already have.
However, you state a lot of good points which I believe should be explored further so that positions and opportunities can be set up when the IC needs more nodes to be onboarded. Please continue to lead public conversations around such possibilities! The IC will need more nodes at some point in the future, and it would be very good if there are opportunities for NPs to onboard at lower expense already set in place and prepared when that day comes. That would be a great boost for decentralization and the future of the IC, in addition to helping to solve some of the funding needs that you have stated so clearly.
I need to clarify that I didn’t propose DFINITY or anyone else selling their servers. There simply are not any existing node servers available to purchase and I understand that it wouldn’t make sense for DFINITY to sell their servers when they have to go buy more.
However, I believe new servers could be created that will comply with the specs. These might be new, but they might also be used. Regardless, this proposal was about reserving the node spot (some people might call this a chit…a short official note, especially one signed by a person in authority, that grants the bearer permission to do something) so they could be filled once new servers are assembled.
It would result in less inflation reduction, but for a good cause. The total number of nodes working for the IC would not change. The number of new node providers would increase, but not the number of nodes.
It seems this discussion could happen now before widespread claims of inflation reduction from node provider remuneration are distributed. Perhaps the inflation reduction would only be 25-30% instead of 33%. Regardless, these are all still good numbers.
Thanks for being upfront about this @katiep. So Proposal: 134282 (at least) should be rejected (posted here)
My perspective is that verification is not about asking the proposing entity if the proposal is okay, and then believing them when they say it’s okay (for this sort of reason, or worse). This isn’t how the IC should work.
If I’m understanding you correctly, there’s no way for me to verify the claims made in the proposals, which means I can’t vote to adopt any of them. I was considering rejecting these proposals in any case, as they do not point voters to where or how they can verify the primary claims of the proposals.
@Lorimer @wpb allow me to make a summary so to make sure we’re on the same page as to the proposals, even for me this is a very long thread to go through. These proposals are related to the four year period of operating nodes, that is expiring for all node providers between Dec’24 and Jan’25. As per this motion proposal from the Gen1 node providers themselves, node providers can continue to operate a max of 42 node machines at a lower reward, and (if they want) sell any excess nodes to other new/existing node providers.
There are two separate cases related to the 18 proposals:
17 of these proposals are related to what @katiep clarified as nodes for which Dfinity is the owner. These nodes are in four data centers (AT1, FM1, SJ1, TP1) and will be partly sold (AT1) partly repurposed.
The proposal 13482 is from 87M, the first node provider for which a proposal is submitted to remove excess nodes from the network so these nodes can be resold or repurposed. However, this proposal contains to wrong data center name (a data center in which they intended to keep nodes), so it needs to be resubmitted.
So the next weeks, we will see more proposals related to this:
proposals of Gen1 node providers with updated node type to reflect lower rewards.
proposals of Gen1 node providers for a new node operator record (only if the node provider decides to redeploy their existing nodes without HSM)
proposals to remove excess nodes that are active in a subnet from the network, so these nodes can be removed and resold/repurposed.
new node provider proposals and new data center proposals if excess nodes are sold to a new node provider that is setting up excess nodes in a new data center
proposals to remove any existing node operator records for NPs that stop operating (which also results in stopping any rewards)
proposals from DRE team (end of February) to remove the original reward scheme from the registry.
Unfortunately this process is not fully automated (it’s complex yes…), so please bear with the team, and all reviews are very muc appreciated and welcome.
The point from @wpb on what to do with any vacant spots is interesting, I suggest to explore the pros and cons further in this thread.
Hopefully this clarifies!
Thanks @SvenF. I know the thread on this topic is long at this point, so I will try to summarize my current thoughts after having re-read everything again.
The nodes in data centers FM1, SJ1, and TP1 are owned by DFINITY, but operated by others. There are a total of 77 nodes in these three data centers. These node operators have been receiving remuneration as per the current 48 month contract, which will expire in Dec24 -Jan25. Since they don’t own the hardware, they cannot move the nodes to new data centers or sell the nodes. At the end of the current 48 month contract, DFINITY is planning to remove all of these nodes from the data centers and repurpose them internally.
Proposal 132553 is the proposal that the NNS approved that defines Gen1 remuneration after 48 months.
Hence, the intent of proposal 132553 seems to be that every opportunity should be given to preserving the number of nodes that are in service. The inflation reduction is supposed to come from the 33% reduction in rewards that Gen1 node providers will receive, not by vacating node chits.
I see a fantastic opportunity to advance decentralization of the internet computer by letting the NNS assign these node chits to people and organizations who have actively contributed to the protocol. I certainly believe that CodeGov should qualify. @Lorimer has also proven his capability and may also be interested. Other organizations that come to mind who might also be interested and have actively pursued concrete contributions to advancing decentralization at the protocol level include ICDevs (@skilesare) and @louisevelayo from Aviate Labs (he is already a node provider, but may not be planning to operate 42 nodes).
Hence, I would like the opportunity to make the case to the NNS to assign any vacant node chits that can be identified over the next several months to these people and organizations. If the node chits can be allocated, then we can start talking about how to obtain the hardware. If DFINITY isn’t interested in selling, then we could build new machines. We could start a capital campaign to raise the funds…or launch an SNS focused on funding node providers who contribute to the protocol, but don’t have the capex to buy the machines.
Letting the NNS stand these people and organizations up through node provider remuneration and allowing them to be independent from DFINITY does far greater good for the decentralization of the internet computer than the overall negligible inflation reduction that will occur if these vacant node chits are allowed to disappear.
Thanks for thinking of me @wpb. I’ve probably misunderstood something along the way, but are you suggesting the community fund raises / gifts these node machines to other members of the community (rather than them purchasing the node machines themselves)? If I understand correctly, the idea is that individuals or organisations performing proposal reviews could then fund themselves by operating these node machines and being rewarded for it?
If I’ve understood correctly, what I’m not sure about is what’s to stop these individuals or organisations from halting their proposal reviews, and just operating the node machines (after all, that’s where there rewards would be coming from). I may have missed a key detail, and I like the intentions of the concept, but I think there would also need to be something holding these individuals or organisations to a commitment (strings attached, which, if severed should result in loss of either the node machine itself, or rewards from that node machine that was gifted to them).