[TEMP CHECK] Revisiting the Operator Marketplace and Fee Models

Proposals
22

Regardless of our opinions of the dynamics, the reality is that there is far more supply (operators) than demand (validators), and no matter our fee policy, it’s impossible to produce sustainability for our current operator set without 1) increasing demand, or 2) extended subsidies (very unlikely).

With that said, if we want to examine this from a theory angle, what you’re describing are standard price discovery dynamics of an oversupplied market, which have been thoroughly studied. There are indeed pitfalls that can cause market breakdowns in certain known specific conditions (e.g., monopolies, unique advantages, etc.), but I don’t see anything about SSV’s market that satisfies those conditions. That’s not to say that there won’t be many operators forced to shut down and maybe some challenging market conditions before the supply/demand equalizes, but I think this is well-understood territory that points to a good (albeit temporarily painful) outcome.

Operators can operate at a loss and try to play games like what you described, but if their motive is profit, they can’t do so permanently. I know you’ve seen markets where this was sustained for a long time (and SSV has been one of them), but inevitably it will correct. Unprofitable operators will eventually quit until supply/demand is balanced. Since it’s already been over a year in this state and the reality of the excess supply is becoming apparent, I don’t think it will take much longer to see change. And we can try to support (maybe even encourage) operators’ shutdown (e.g., I know some are concerned that they have some validators running on them, and we should help them to gracefully exit and mitigate customer impacts).

The only way this won’t resolve to ~equilibrium is if there’s somehow a massive endless stream of new operators that foolishly join when few are profitable, attempt to undercut, fail, and leave, creating a perpetual cycle of failure with no room for anyone to grow. But we don’t seem to be seeing this pattern now, so there’s no reason to believe it will happen. And we have enough tools to steer away from that outcome: better cluster selection tools that emphasize long-term performance and trust, better operator data, simply advising new potential operators that they shouldn’t join, or even temporarily closing public operator registration if it comes to that.

2 Likes
23

Forgive my further rambling to flesh this out and extrapolate along our situation…

However, with the above said, if a large number of operators shut down right now, it would be bad because:

-There’s no way for operators to contact their stakers to warn them.

-It’s difficult to swap operators, especially if it’s already been done and keys need to be regenerated (plus the current 50+ day exit/entry round trip).

Lots of customers will be upset, so I think we want to avoid this if we can. As I said before, if we feel that we need to buy ourselves more time to avoid a mass exodus of operators before we can mitigate the customer pain of operator shutdowns, then that’s an option… a last-resort high-cost option, but an option.

But after we’re able to remove those two issues and make operator shutdowns insignificant to stakers (as we plan to), is it really a problem if operators quit? We don’t need every operator to be successful, and I don’t think the DAO has the responsibility to ensure that they are. We just need enough to match demand, which is probably <50% of what we have now. This can grow more later as demand for public operators increases (which we should be aggressively working towards).

I know all this might seem harsh, and I know we could also look at this from a perspective of pain minimization. If the community feels that it’s important to provide one-time support as goodwill, then I won’t argue. I actually think it’s probably the “right” thing to do (and could buy us time to put operator shutdown impact mitigation measures in place). But I’m hesitant to support such things in practice because 1) we need operators to quit and probably shouldn’t enable them to stay, and 2) additional spending is hard right now.

Along these lines, I also don’t think we should enact a minimum fee or similar mechanism for any type of long-term union-like or employee protection purpose, because:

-I think we should see starting an operator as pursuing a business venture. Success isn’t guaranteed, there is risk, and this is simply the game we’re all playing. It’s each of our decisions on whether we want to participate. You can argue that the market has been “unfair” due to the inability to change fees, token price drop, lower-than-expected demand for public operators, etc., but from here forward, I think we all understand what this is.

-There is no exploitation here, just competition and market immaturity. Once supply/demand balances and we fix some of our major issues, it should be a pretty nice gig for a lot of people, even at small margins. For example, even a ~$1000 profit per node per year is meaningful for many, especially considering the required labor is like, what, 15 minutes a week per node on average? (~$77 / hr)

-Good (enough) operators are abundant and it’s unlikely that we’ll ever have a significant shortage. Let’s be honest… running an operator is getting pretty easy with how mature the technology has become. Thanks to the amazing tools that have been developed by the greater community, most aspects of the job are automated out-of-the-box, and what’s left isn’t time-consuming. This trend will likely continue. The main qualities of a good operator are moving towards trustworthiness, devotion, and responsiveness, rather than advanced technical expertise or experience. I think there’s a huge number of people that meet this criteria that would fill this role for a high likelihood of a meager profit.

-A low sustainable “market rate” operator fee is good for the network. Lower operator fees improve competitive advantage and greatly increase the upper bound of our network revenue. I think it’s also good for Ethereum as a whole.

I think we should also remember that DVT is a new disruptive technology in this space, and we should not hold onto the past “industry standard”. DVT enables a new node operation paradigm that expands the accessibility of the operator role, reduces operator costs, and increases decentralization. What we’re facing now is economic pain from the transition, mainly because the engineering isn’t finished and the market is still broken for several reasons. We simply have to look forward, finish the work, and advance past these issues.

2 Likes
24

Thank you @fod and @BumpyTale for this debate of ideas surrounding the public SSV marketplace fee. That could be a great case study for a tokenomics class in a university. :laughing:

In a free market, nothing is guaranteed, some players go under, others take off. That cycle of winners and losers is what keeps things efficient and pushes innovation. If you try to protect everyone from failing, like by setting a minimum fee, you break that process. The “invisible hand” only works when weaker players can drop out and stronger or more competitive ones rise.

That being said, I totally agree with @fod. The problem here is not the absence of a minimum fee, it’s the restricted mechanism to increase operator fees.

I think there should not be a limit on changing operator fees. There should be a mechanism that gives stakers enough time to react if an operator increases its fee, let’s say one month so validators have enough time to adjust and avoid liquidation. This cap on fee increases goes against basic concepts of the free market.

SSV public operators are probably the only staking providers that allow stakers to consolidate their validator into 2,048 staked ETH and only pay for one validator (not pay by Effective Balance). Lido, EtherFi, StakeWise, Swell, etc. are still spinning up multiple standard 32 ETH validators, not big consolidated ones!

So yes, it’s kind of urgent to fix that amazing hack for stakers by setting the Operator Fee based on a percentage of Validator Earnings, not per validator. At AXBLOX, we have 3 consolidated validators on the public marketplace. We tried to increase the fee over the last month and a half, but starting from 0.3 SSV/year it’s ultra slow because of the max 10% per 2 weeks increase.

1 Like
25

About that:

Maybe my message has not been read in the private vo-committee channel on Discord, but I spent quite some time analyzing the SSV official documentation and here’s what I found (I personally did not use DKG until now, always KeySplitting method):

With SSV’s DKG, the validator key is created so nobody ever sees the full private key. It only exists as encrypted shares across the operators. When you need to swap operators, you run a reshare:

  • The private key stays the same, so the validator identity doesn’t change.

  • Each operator gets a new encrypted share of that same key, old shares are thrown away.

  • A quorum of current operators must cooperate so the secret never gets exposed.

So yes, DKG-generated keys can be re-shared on the SSV web app to swap out operators on the fly without exposing the validator key. It is one of the biggest advantages of DKG. If the validator used the older KeySplitting method (the validator had the full key first then split it with the operators) then he/she can still change operators by manually re-splitting or re-onboarding with DKG. But re-onboarding with DKG defeats the whole purpose of DKG. The whole point of DKG is that the private key never exists in one place. If you already had the full key once, migrating it into DKG later won’t give you the same trustless guarantee, the validator key has already been exposed. Better to start fresh with DKG from the beginning.

Let’s add to the DKG approach:

With DKG, the validator/cluster owner never has the signing private key. Exits and operator resharing require a quorum of operators (≥ threshold). Even if the cluster owner loses quorum for example, 2 out of 4 operators disappear in a 4-operator cluster, the validator is still safe. Thanks to the Pectra upgrade, you can now exit a validator and withdraw the stake directly from the withdrawal key. That means you’re never stuck, and this makes DKG an obvious choice.

1 Like
26

I don’t agree with your arguments, but I think I’ve made my perspective clear with the prisoner’s dilemma example. I’ll leave it here for the community to weigh in and decide how we want to approach this.

1 Like
27

I think a clarification is needed here:

Resharing changes the shares, not the secret. In SSV’s DKG, resharing regenerates new shares of the same validator key. If ≥ t former operators kept their old shares, they can still reconstruct the key. “Old shares are thrown away” is an operational assumption, not a cryptographic guarantee. To fully revoke past operators you’d need a key rotation (new validator).

1 Like
28

Thank you @BumpyTale for the clarification!

Yes, resharing only re-splits the same validator key, so if old operators kept their shares they could still help reconstruct it. The protocol assumes they discard them, but that’s an operational risk, full revocation requires a new validator key.

Collusion risks

Please do not change more than 2 operators in a cluster. Each set of generated shares will always be valid when their signing threshold is met (e.g. 3/4). To reduce the risks, it is advised to not change more than 2 of the validator’s managing operators when changing its cluster.

Update Operators | SSV

There’s no easy fix to that other than a protocol-level change maybe something like adopting an advanced secret refresh protocols and aligning them with Ethereum’s validator key structure. Which is not (I believe) in the SSV roadmap.

That being said, the recommended Idea summary that @GBeast wrote about this Temp Check i.e. “Cluster Updates - Swap out operators for new operators seamlessly” can be removed in my opinion.

29

@BumpyTale I don’t think a minimum fee fixes the problem. The “prisoner’s dilemma” idea only works if price is the only thing that matters. But in SSV public operators market, stakers choose operators for lots of reasons: performance/uptime, reliability, reputation, client, geo diversity, communication and overall trust. Fees are one variable.

Also, swapping operators takes work and has risks (collusion risk like mentioned above) so stakers don’t jump ship the second an operator adjusts its fee. That means operators can stand out on quality and trust, not just on being the cheapest.

Free markets always have some price pressure, but here competition happens on many fronts. A minimum fee just cuts off natural competition instead of letting good operators prove their value.

30

Am I the only one having this issue on this Discourse forum lately? I can’t scroll down on long messages I am editing. It’s pretty annoying, I tried different browsers but no luck.

31

I had the same problem. I had to zoom out 500% just to see the “Submit” button.

1 Like
32

I don’t believe it’s true that most stakers choose operators for those reasons you stated. In my seven years in the staking business, I observed that there are only two variables on which most stakers choose operators, which are a) being a top operator in terms of TVL (when others choose this operator, it must be good) and b) low fees. Usually b) kicks off the race to the bottom, as it’s the only variable on which operators can compete, when they are not in the fortunate position of being backed by venture capital and/or having a strong institutional customer base - and thus TVL.

BTW, the prisoner’s dilemma is not just an “idea”. It’s a well researched conceptual model in international politics and social sciences.

1 Like
33

I’m weighing in on one more point and then I keep silent:

I believe the DAO shouldn’t be governed on whatever worldview individuals have, it should be governed on the grounds of what is best for the protocol and its long term development, and thus, a carefully balanced approach is warranted.

The SSV DAO shouldn’t be a testbed to try out one’s rather extremist, libertarian ideas of a “pure free market at all cost”, which would wipe out any semi-professional staking business and only leave hobbyists (who don’t care about profitability) and hyperscalers (who can absorb losses).

The extreme libertarian ideas I’ve read here (and elsewhere) treat operator attrition as a feature, not a bug, which, in my opinion, endanger the very ecosystem we’re trying to build.

2 Likes
34

Hi @BumpyTale

Hope you’re okay if I keep debating with you, ultimately it’s in the interest of the SSV community. :grin:

You’re making some really compelling points here!

But, :sweat_smile: the prisoner’s dilemma doesn’t always apply, it’s only about clear cooperate vs betray choices. On the other hand, the SSV marketplace kind of fits cause operators can raise fees for short-term gain (betray) but if too many do, stakers leave and many loses. Still, it’s not a perfect match with many players, reputation and long-term play involved.

I agree when you write “a carefully balanced approach is warranted.” In that spirit, why not, as a cautious first step, allow more free-market fee adjustments (up to a 100% monthly increase as proposed) to see how the invisible hand of Adam Smith operates?

It’s possible that we observe that operators follow their own self-interest in a freer market, which could unintentionally benefit the marketplace as a whole. But if after some time, we notice “operator attrition”, then we could revisit the idea of a minimum fee.

1 Like
35

I’m very curious to see how the natural market dynamics will play out if operators have more freedom to adjust their fees.

I also think it would be interesting to push these Temp Check suggestions into one proposal focused on improving the Operator Marketplace and Fee Models then on a separate proposal only the suggestion on setting a minimum fee. This way, we would see the community’s true position on the minimum fee. Let’s not do what American lawmakers often do, pack lots of different measures into one giant bill so it’s harder to reject, even if people disagree with parts of it.

In the end, it’s the SSV holders and delegates who vote on the proposal, making it a democratic process.

1 Like
36

I don’t want to continue going back and forth forever, but I feel such comments warrant some clarification. Then I too will leave this issue up to the remainder of the community for now. But despite my strong opinions, I hope it’s apparent that I’m sincerely considering your concerns.

one’s rather extremist, libertarian ideas

First, I don’t think the encouragement of solo stakers/hobbyists is extreme. It has been a core tenant of the ETH ethos and a celebrated focus of several communities and protocols, like Rocket Pool and EthStaker. Much of the reason for my views is to give such operators an advantage over larger players.

Second, I don’t think the search for mechanisms to increase revenue and reduce costs is extreme. It is a standard element of any business, especially in an environment of significant competition before we have reached a point of financial stability.

These are straightforward goals that I think almost everyone here would agree with in some capacity, and DVT enables dynamics where both goals can be achieved synergistically. It’s just up to us to design the best solution along this obviously-beneficial vector (not necessarily exactly what I am suggesting).

which would wipe out any semi-professional staking business

I am certainly not proposing the removal of professional operators, they certainly have a place in SSV, and I disagree that a free market would wipe them out. However, I think it’s necessary to be realistic about both DVT as a technology and the evolution of ETH staking as a whole. Because the inherent advantages of DVT put any good operator on nearly the same performance level as professionals, if professionals feel they deserve a higher fee than hobbyist ones, then simply having high performance and ~100% uptime is no longer sufficient.

But there are many things an operator can do to provide premium service. For example, things like multi-client setups to reduce risk, using Weighted Attestation Data to maximize rewards (https://ssv.network/blog/technology/maximising-ssv-validator-rewards-with-wad/), or providing an exceptional level of customer service/updates, etc. We also have the option to highlight these advantages within the protocol and our customer-facing tooling if we choose to encourage users to value them.

treat operator attrition as a feature, not a bug

I treat it as neither. It’s simply the reality of our current market that we must all accept (or work hard to change). It is undeniable that we have far more public operators than can be supported by our current customers choosing them. I still feel strongly that our primary effort (beyond what’s described in the new temp check) should be to attract more customers to public operators by increasing the value we offer. The only other option is forcing many operators to shut down.

minimum fee

I spoke to several more people about this issue, and I see some additional value of a minimum fee for protection from the possibility of bad actors or those that try to amass large numbers of validators with motives other than profit. I think these are small risks but risks nonetheless. However, I still feel strongly that there is a lot of value in giving hobbyist operators a way to exercise their cost advantages and in trying to encourage a low “market rate” for operator fees to help the protocol’s cost/revenue (again, we could likely double our revenue or more).

I would support a minimum fee around 0.25% per validator share (maybe slightly higher), targeting the cost of a competitive hobbyist. At this fee level and current market conditions and using 32 ETH per validator, 50 validator shares would net ~$500 per year, with a very successful operator attracting 500 shares earning ~$5000.

…But if we implement this, we should strongly consider a mechanism that gives advantages to new operators, to help them get their first few validators and start to build trust. And I think having a mechanism to encourage or incentivize spreading validators across more operators would go a long way toward addressing issues of over-concentration and making more operators profitable.

Regardless, let’s push towards solutions for the more straightforward and pressing issues (proposed in the new temp check). We can keep discussing this, but it’s clear that the other problems should be fixed ASAP and that we have consensus on those. And I think with at least some of those problems behind us, we’ll gain more information to help guide us with what’s left.

2 Likes
37

Well said @fod this debate is still a work in progress and the discussion should continue.

On that note, AXBLOX is planning to do exactly that:

Our public operator ID 423 just got 2 new 32 ETH validators with you @fod in the cluster, the SSV Public Network is still alive. Let’s keep pushing it forward!

1 Like
38

I believe this can solve the fee volatility challenges

2 Likes
39

Big tokenomic shift!

If fees shift to ETH, how do you see SSV’s utility being maintained beyond yield and governance? Could there be hybrid mechanisms to preserve some native token demand?

And if SSV starts distributing ETH yield, do you think governance participation will increase or will it concentrate among large stakers chasing yield?

40

There are a few possibilities I can think of, i’m still trying to finalize it in my head.

I actually think this will create much more wide spread distribution of the token as it opens the door to many more types of holders

1 Like