Pre-Proposal: Relay Driven Inflation (RDI)

Attributes

Author(s): @crabman , @TheDoc , @mess and others [private]
Parameter: RelaysToTokensMultiplier

Current Value: Variable set by the SER Proposal https://forum.pokt.network/t/pup-30-sustainable-emission-reduction-ser/3937. Currently approximately causes 10-12% inflation or 700k-762k USD per month

New Value: Variable, decouples inflation from being supply based to relay based. Pokt token are minted proportional to the index price of a relay in USD and the current price of pokt. As of today, this would cause an effective inflation of 4.55% or 240k USD per month.

We also introduce a new parameter that can be modified to keep this proposal adaptable to changing circumstances.

RelayPricingIndex: Monthly price for an average daily relay.

Initial value:
RelayPricingIndex: 0.0002 per relay [ daily for 30 days or 30 relays ]

Calculated using Infura Pricing based on the Growth Plan (Pricing | Infura)
$1000/5m = 0.0002 per relay

Summary:

Although all our previous inflation measures, including SER reduced inflation, they were detached from the ground realities. As a result, inflation remained high, and the protocol does not appear to be in any position to absorb even half of the minted POKT. Not even towards the end of the SER cycle where we expect inflation to be the lowest (based on current pricing).

We believe this can be detrimental to the price of the token and will discourage entry of new buyers or those who simply want to hold the token for the upside.

Please note this is intended to be a temporary inflation mechanism. When V1 comes about we should restart some sort of bootstrap to attract node operators again because at that point we will have the necessary ingredients to increase our relays, especially the paid ones.

Abstract

To balance out inflation caused by SER, at the time of writing this proposal we need sales of 762k USD every month. Using RelayPricingIndex based on Infura, this would mean we have to scale our relays to 762k/0.0002 = 3.81 billion paid relays daily

This is a stretch for the following reasons:

Data Integrity: At this time, we have no data integrity built into the network. We do, however, check for the viability of chains behind the nodes at the cherry picker level by making sure their block height is current. We will not get customers who build critical dApps or trading tools to use our service let alone pay for it.

Exhibit A

Link to Discord Post

Architectural limitations: We cannot go beyond 3B relays without tweaks which are still being worked on. No hard deadlines have been given. We believe we need to scale to atleast 5B relays including free relays to realize 3.81 billion paid relays.

Exhibit B:

Link to Discord Post

Challenge for our Sales Department

Selling ~4B relays is a very challenging task even if we somehow work the above two.

Exhibit C

Link to Discord Post

SER gives a Moving Target to our sales Team.

At the price of $0.04 USD per POKT, we need 762k USD sales monthly. What if price doubles or triples?

Note: Gateways costs are not included.

Motivation

This proposal Vs SER.

  • Above shows the difficulty of getting our sales to the levels required by SER. Also shows inflation is not really going down fast enough.
  • Doubling the price, doubles the sales number while this proposal keeps it the same.

  • This proposal helps price recovery by keeping the USD value of pokt emitted the same while decreasing inflation.

Additional benefits of this proposal:

Makes POKT token desirable to hold:
Today because of high inflation, unless you stake your pokt, your token will be inflated away. This is one of the reasons why our order books have such bad liquidity. It does not help that we also have a less than ideal staking mechanism.

Gives power back to the node operators who own their hardware:
We should be encouraging node operators to use their spare hardware (eg. home servers) or invest in buying hardware to run nodes and chains. We have so far been enriching cloud providers. Node Operators who own their hardware can operate chains at a much lower cost and will not mind the low inflation.
Moreover, since they know inflation is low, they will not be inclined to stake with a node operator to get “full rewards”. They will be happy to get some rewards running the top 2-3 chains on their home server.

Saves us inflation for when we are at the top of the game
This proposal will save us about 95m tokens which could potentially be used to bootstrap v1.

Specification

RelayPricingIndex: 0.0002 per relay [ daily for 30 days or 30 relays ]

Average Daily Relay: 1.2B (Current)

Using above, set RelaysToTokensMultiplier to target an emission rate, In the example above that will be 200k tokens daily.

We hope that the sales team can generate revenue close to that number and buy pokt off the market to balance out the inflation.

Start of this proposal, inflation must be cut drastically to get to levels provided by this proposal. It can be cut in one month in the following way:

Dissenting Opinions

Node Operators will not have enough time to adjust:

Node operators will get one week before the beginning of the new month to reduce their operations. A month following that week will have decreasing inflation. Commercial node operators also have a “treasury” which they can use to navigate for the time being.

Node Operators can also reduce the number of chains or regions they serve to reduce their expenses.

Node Operators unstaking will be bad for the network:

This is game theory, if half of nodes unstake the rewards will increase for the other half.

We are also still massively over-provisioned. This is actually good for the network. This forum post by poktblade is an interesting read.

We also have a very strong Altruist network to carry the load if required but it’s unlikely we will get to that. We have 20k nodes staked, network can easily run on half of that.

Unstaked users will dump their tokens.

Low inflation should ideally encourage them to hold the tokens, besides many node providers are now offering 0, negative or very tiny yields. This along with high inflation is eroding the value of staked pokt tokens. Surprisingly, the pokt tokens are still staked to those providers so we have no reason to believe tokens will be dumped.

Copyright

Copyright and related rights waived via CC0.

7 Likes

Thanks for posting. I’m out all this week but look forward to reviewing this next week when back online

1 Like

I am not an tokenomics expert but I think this is definitely a step in the right direction regarding inflation and $POKT price. :100:

4 Likes

My initial thoughts:

I think tying the token emissions to a more tangible metric like paid relays is sensible.

My supposition is that this will still cause some node runners to close up shop, but more in a rip-the-bandaid-off way rather than the slow, painful grind we have been experiencing up to now.

It seems like in practice, this proposal deincentivizes geomeshing, which may not be a bad thing.

1 Like

Thanks for this post @crabman! Still munching on this, but got a question:

I believe this is inaccurate. To get a price per relay, you need to multiply the 5M by 30.4 (to represent the average days in a month). Infura’s plan is 5M per day, so that technically equals around 152M relays a month, which the $1000 covers.

I believe the more accurate price per relays would then be $1000/152m = 0.000006578947 per relay

Does that track or do I have something off?

2 Likes

Thanks Crabman for putting up a thread with your proposal. I’ve been following this discussion in the unofficial channels and it was time to have a proper place to talk about it.


TLDR

While the concept is valid and more natural than SER, I don’t think that it is wise to implement it. The idea of balancing emissions is not correctly justified and the bounding of minting to relay volume without a paired burn mechanism was the cause of the initial surge in inflation of the Pocket Network.


Full Answer

I think that there are some problems with this proposal, I will try to separate them into categories.

Bounding minting with relay volume

I like the idea of coupling emissions to relay volumes, this is the way Pocket is supposed to work. I have mentioned a few times that the decoupling of relay growth from income growth is not natural and creates a dissociation between the sales objectives and the node runners objectives. However I think that we are not ready to bound emissions to relay volumes again (like before WAGMI).

Over the last year the number of daily relays has increased from ~800M to 1.2 B, this is an increase of 50%. We expect this increase to be even higher this year if sales objectives are met and removal of free traffic does not compensate all sales.
The scenario of relay growth is not taken into account in the provided tables.

Since there is no burning mechanism, the emissions were decoupled from relay volume to avoid this situations, where increasing the relay volume (something positive) results in uncontrolled emissions (something negative).

Inflation Balancing

You write about “balancing out inflation” and how many sales are needed to compensate emissions, the problem is that the sales are not network profit as @Caesar points out in his blog post. PNI will never buy back and burn 100% of their sales incomes because that’s not how they work (reference needed).
Until burn is activated and each relay done translates into burning, there will be no “balancing” of emissions. This is true for SER and would be true for RDI (this proposal).

Relay Pricing

Setting the price of the relay has lots of implications, it is not only fixing a cost, you are in fact fixing the real exchange rate of $POKT to a value controlled by a third party (in this case Infura). This might sound weird here but we will be posting shortly (next week?) something about the macro-economic mechanisms of the Pocket Network, and the relative price of a relay in the Pocket ecosystem to the price of a relay outside Pocket is a fundamental variable that we should be able to control.

Also, the PNI portal page states that the cost of a relay is U$D 0.000007456, that’s 26 times lower than the proposed value. This means that any client of “Pay as you go” plan will have a massive increase of their price tag. As you say, Pocket Network is not perfect and we have issues of data integrity (that hopefully will be solved in portal V2), will the clients accept this or they will just fly of to Infura for the same price tag?

Also a note on the free relays, they are not part of the protocol, each relay must be paid for (meaning that each relay will burn POKT stake in the future). Free relays can be given but someone will have to pay for them, PNI, the portal operator or the community (through DAO).

Calculations - Study cases

You provide tables (please share the google sheet if you can) for a single study case:
What if the price goes up? (ceteris paribus the rest of the variables)
You show in this case how the SER would mean that more U$S are “printed” (I disagree with presenting this figure in U$D tho), on the other hand the RDI would keep that value constant.
But what about the other variables?

  • What if the price stays constant or goes down while the number of relays goes up? (current scenario of the Pocket Network)

I feel that we would need to observe other cases, not just the “happy path” of the proposal.

Network Cost and Node Running

I feel that we talk about network cost and node running costs very lightly. I would like to see some number behind the claims like:

In my experience this is not true, the Pocket Networks is already in low/none-profit for many node runners. These claims often overlook the human resources part of the equation (which are fundamental to the Pocket community).

I also don’t understand why this will be true:

If a node provider gives you 75% of 12 POKT a day you will earn 9 POKT, but if you stake 2-3 chains in a single region with no-so-good QoS you will be below 9 POKT (you can check this on POKTscan).

With lower emissions the first to die will be the smaller ones, big ones will only feed on the increased share of emissions that are going to be released by smaller NR who capitulate. You are encouraging centralization.

This is bad for the Pocket Network, this will drive us back to pre-GeoMesh QoS distribution (some more info in the GeoMesh thread):


(and this is an outdated image, from GeoMesh release, now we are even better)

The only way to remove geomesh without impacting the QoS is by regional staking, and that would come in V1. If you want to help small node runners and optimize hardware usage, I have already proposed more restrictions for V1 here.

This is a common misconception, we cannot measure over-provision by counting Pocket nodes. The quote is very outdated, you can ask @poktblade if he still thinks that a Pocket Node is something to take into a account when measuring the cost of a node running operation.
All arguments based on this are not valid. The network can run with a single 15K node for each real provider, but that will not happen because we are greedy, secretive and internet annons.

7 Likes

0.0002 is a monthly basis price.

2 Likes

As a co-author of this proposal, I’m obviously in full support of this proposal!

There are some more things that should and will be added before this proposal goes into a final proposal stage, which is making clear what are the exact outcomes of this proposal and reflections on inflation rate in various situations such as the following:

  1. relays go up
  2. relays go down
  3. POKT price goes up in USD
  4. POKT price goes down in USD

Also, it should be crystal clear that this proposal is supposed to be in place as long as POKT has v0 on mainnet and this coin economics stops and should be replaced with a better one the moment when v1 gets activated.

The immediate result of this proposal upon implementation is basically where SER proposal will be at the end of its cycle. The only difference is that this proposal saves POKT investors from months of pain due to the imbalance of current inflation rate (newly minted supply of 700-762k USD per month) compared to paid demand currently being locked in on a monthly basis by a POKT Biz-Dev/Sales team. Imbalance mentioned in proposal has a corrosive effect on a price throughout the entire 1 year period until v1 goes live. This proposal fixes it by instantly reducing the inflation rate to a healthy equilibrium level of freshly minted supply and paid demand. Because of the prevention of excessive coin emission (inflation rate) for a whole upcoming year, this proposal saves the POKT price at levels most likely way higher than levels where SER would lead during the full 1 year of imbalance in place.

Purpose of this proposal is to stabilize the tokenomics, find a new ground price from which POKT can only start going up and be able to survive until v1 gets finished and implemented on mainnet, starting a starting a whole new chapter of Pocket Network.

5 Likes

Relay Pricing

Infura pricing is just used as an index since it’s the gold standard. Happy to average the relay pricing out with other reputable providers if you know any. I picked infura only because they are the oldest and they price their relays per quantity (same as us) and not by compute units.

You also misunderstood the relay pricing. 0.0002 is the “monthly price” for an average daily relay.

Infura charges $1000 per month for 5M relay daily. So that’s how that figure is calculated

$1000/5m = 0.0002 per relay

0.0002*1.2B = 240k USD emission monthly

Price per single relay comes out to be $1000/152m = 0.000006578947 per relay as calculated by @shane

This is 12% less expensive than PNI price per single relay.

image

To customers our strength over infura is the variety of chains, especially archival, and more reliability against outages. Data integrity is our weak point.

Bounding minting with relay volume

All the relay growth were free relays, as PNI is shutting down free relays I expect growth of relays to be slow but compensated for. Our highest daily relay was 1.5 billion. Also from the graph, overall relays have not changed much since August of last year. Always going back down after touching 1.3 billion.

Link to Discord Post

image

Btw here is the emission at 1.5b, still not bad

image

A major problem with SER is, we can get to a point where USD emission will not be able to support the existing relays or growth of our relays. This is going to be disastrous for the network. Our bills and sales are in USD so should be our emissions

However, it is only slightly better for RDI, we can continue to run our network at least although inflation will go up. Overall, I agree it’s a problem for both, and we should not get to this stage which brings me to the next point.

Inflation Balancing

Nowhere in the proposal I mandated PNI buys off pokt from the market and burns it. I only “hoped” they could. I understand they have other expenses. This proposal only brings the finish line closer. PNI being able to buy/burn 50k-60k USD worth of POKT will make a greater impact than if they did it at 760k USD emission.

Also, the situation of pokt being at $0.01 and 1.2B daily relay needs to be avoided at all cost and we need to think ahead and start burning some from now or reduce the free relays. This applies to both SER and RDI but again that’s for later and it’s not part of this proposal

Network Cost and Node Running

The human cost comes from managing clients (bad staking mechanism) and forcing node operators to run all 15 chains(in three regions). Think of a scenario, let’s say I am running eth validator at my home, maybe I should also provide RPC to make some income from pocket network.

Those are the people we should target. They own their hardware, so they realize the full APR pocket network is providing instead of giving a chunk to ovh/hetzner etc.

Smaller ones who own their hardware will not die.

No node operator will give out 75% of rewards if this proposal goes through.

Regarding geomesh, we will still have that. Large Node operators can maximize their rewards vs. cost if they choose to run top 5 chains in all regions. Since the chains are the same, they will reduce their human cost and running a limited number of chains will reduce hardware cost.

This will leave out bottom chains for small node operators. We should also just let the pocket economy dictate if they want to run another chain in the same region or the same chain in a different region.

I don’t want to help small or large node runners; I want to reduce network costs. If a small node operator can run the chain on their home server for cheaper that’s good. Or if a large node operator owning hardware in colocation data centers can serve for cheap that is also good.

We should not discuss decentralization at this point. Let’s make this network work for customers and investors first!

Overprovisioning

We are overprovisioned by any metric you look at. Whether it is number of tokens staked or nodes. If you disagree with that, I have another metric for you. We currently have 62 distinct domains running ETH chains. 62 Eth chains for 500k daily relays tops? If I add in geo mesh that makes it worst. If you do this for chains having 100k relays and less, it is even worse.

Domains may share chains, but we will never know, but I assume just from experience that not many people are sharing, because with non-custodial staking you could simply delegate your tokens in a safe way. Also, CC is not live yet.

Let me know if I missed anything, or you have some other concern.

2 Likes

Unfortunately, this is a hard but necessary step. Node Operators especially big ones have benefited the most from inflation especially by using a version of LeanPokt and Geomesh before they were released to the public. It is time to end the party.

Small Node operators can still survive if they own their hardware and they should be able to run a couple of chains to get decent rewards. Low Inflation means you don’t have to chase the best returns to compensate for devaluation of your tokens.

1 Like

See my response to Ramiro

1 Like

Hi Shane, the relayPricingIndex represents the cost of roughly 30 relays or 1 relay every day in a month.

It is basically 0.0002 relay per 30 or 30.4 relays so your price of a single relay is correct. Either way it makes very little difference to calculations, and our target emission setting process is also an estimate anyways.

I will make this clear in my proposal,

Thankyou

1 Like

I am in support of this proposal while v0 remains the primary gateway in production. I do agree with some of Ramiro’s points and appreciate his well articulated response. I acknowledge that economics is a counterintuitive domain that requires qualified expertise and experience to optimize, as well as agree with the general sentiment that SER, given current market conditions, will lead us all to ruin if something drastic doesn’t change.

I do also believe that we need to optimize the network to favor node runners who put in the time and effort to improve QoS for the network as a whole, as well as the net return for clients who want to invest in our ecosystem. I’m not qualified to fully understand psychological incentives and economics, but I do think we are wasting money inflating all our bags away to pay low quality node runners who act as dead weight to the rest of the ecosystem.

4 Likes

A complementary point: the users can hardly consume 100% of 152M, so the actual cost would be higher than the theoretical price. On the other side, PNI’s price is based on the real usage. Considering this, probably the real cost of using Infura is a little bit higher than PNI’s price

2 Likes

Yes great observation!

Plus PNI offers 250k free relays vs. infura’s 100k free only non-archival relays.

Using Infura’s pricing to mint pokt for each relay served by our node operators is actually quite generous.

1 Like

Thanks for the answer, I will respect the topics:

Relay Pricing

Thanks for the clarification, now the price makes much more sense.
I suggest that you change the initial post to reflect the value of the relay to the correct one. The notion of price per “daily relay” is confusing and should not have the units [u$d/relay].
The actual value of the relay in u$d for the Pocket network client will ultimately be selected by the portal runners (currently only PNI), the price in POKT of the relay is what the DAO should control. Anyway this discussion is not really central to this topic and I think that the price that you chose is a good starting point. Lets assume it is fair and correct.

Bounding Minting with Relay Volume

Its fair to think that growth was from free relays, however I’m a little more optimistic on the sales team. I will explore what happens if we reach 3B relays in a year.

As I understand you are bounding emissions to relays AND the ecosystems revenue which you obtain from PNI sales (I disagree with this, but lets move forward). If sales go up, the minting should go up. You mention that minting 240K u$d worth of $POKT per month will put us in balance. So, if PNI doubles its sales we should mint more? (I’m not clear on this). This seems to unbound the minting from the relays again, as revenue is proportional to relays, you cant have both… I will explore both scenarios (bounding minting to relays and to revenue)…

Now, here are the results from my “unhappy paths” for this proposal. You will find the calculations here, please let me know if you find any errors/miss-interpretations.
All scenarios share:

  • Time frame: 2023-04-01 to 2024-02-01
  • Initial $POKT price: 0.04 u$d
  • Initial relays per day : 1.2 B
  • Initial $POKT supply : 1.58 B
  • Linear interpolations between initial and end values.

Price Down, Relays Up (current scenario, bounded to your projections)

  • Final $POKT price: 0.02 u$d
  • Final relays per day : 1.5 B
  • Bounding minting to relays
    Result: RDI produces more inflation than SER after January.

Price Constant, Sales Up (probable scenario)

  • Final $POKT price: 0.04 u$d
  • Final PNI revenue: 720 K u$d/month (3 times higher).
  • Bounding minting to revenue (with more revenue we can mint more)
    Result: RDI produces more inflation than SER after November.

Price Down only

  • Final $POKT price: 0.01 u$d
  • Relays and PNI revenue constant
    Result: RDI produces more inflation than SER after November.

Price Down, Sales Up

  • Final $POKT price: 0.01 u$d
  • Final PNI revenue: 720 K u$d/month (3 times higher).
  • Bounding minting to revenue (with more revenue we can mint more)
    Result: RDI produces more inflation than SER after August and clearly exponential, pre-WAGMI again…

Price Down, Relays Up (limit case)

  • Final $POKT price: 0.01 u$d
  • Final relays per day : 3 B
  • Bounding minting to relays
    Result: RDI produces more inflation than SER after August and also clearly exponential, pre-WAGMI again…

Inflation Balancing

According to PNF post (more precisely the document they attach in the post), the current PNI revenue is:

So, from the 1.2 B they only get paid for 200 M and they are committed to buyback only 60% of their revenue [citation needed]. Their revenue that can actually act as buy pressure for $POKT is actually 27.2 K u$d/month according to my calculations (see gist). If we modify your proposal to match this revenue level and keep everything else constant, we would only mint ~22 K $POKT, this is unbearable (IMHO)…

Network Cost and Node Running

We all agree that using hardware that has no cost, like blockchain nodes primary used for other stuff or independent node runners is the ideal situation, however the Pocket Network current incentives are not aligned with this and this proposal will not help that.
I fear that small node operators who do not care about the scaling or the price of POKT probably do not care also for the QoS or data integrity of their operations. But I fear this is a subject on its own and has little to do with this proposal.

Or they can choose to chew more from their clients and use their size to leverage the reduction in incomes from this proposal. They will feed on the smaller ones that capitulate. This has happened before with some node runners charging up to 99% rev share…

I disagree with this, we should be able to survive while keep the decentralization alive. Otherwise lets just pay CC for all the relays, they can handle them.

Over-provision

Its not the same to count Pocket Nodes than counting Blockchain nodes. If you agree with this (as I seem to understand from what you write), you should make this clear in the proposal and remove all mentions to Pocket Nodes in regard to over-provision to avoid confusion.

Yes, we are over-provisioned and we should try to reduce that, but again this proposal is not the right way to do it for other reasons that I mentioned earlier in this post.

5 Likes

RDI is intended to produce just enough inflation at market rate of a relay to keep our system going. Nothing more and nothing less.

I already listed out the reasons why 3B is not possible with v0 in the near term, I will not repeat here, but let’s assume we fixed all our problems.

Assume we are doing 100b relays daily, but the price remains the same, SER will continue to decrease inflation.

Stop and think here for a second about what will happen? Node operators are being asked to do more work for little reward. The result is we won’t be able to support 100b relays and system collapses. Similarly, if token price keeps decreasing and relays stay at 1.2B per day, SER will eventually kill the project.


RDI VS SER: At 126k USD System will fail to support 1.2 Billion daily relays

Flip Side, we have been rewarding node operators a lot more than the work they have been doing and if SER continues, we will continue to do so at this rate of relays.

This is a fatal flaw in SER.

Now let’s go back to the situation where token price remains the same, but we are doing 100b relays daily. Although possible, this imbalance is really a contradiction. We are saying our system is working great, customers are paying, and APR is high to node operators because of RDI but our token price is not increasing? I hope you see that we should not be in this situation for long from an investor’s point of view. The likelihood is high that token price will catch up to bring the APR down.

The buyback you mentioned may help at lower prices of pokt with SER, but as many of my plain tables show, at higher prices they will be a drop in the ocean because if price goes up, SER will cause higher USD inflation than the buyback amount. High USD inflation will put a reverse pressure on the token causing it to go back down. This makes the token unattractive to hold.

Regarding overprovision, it does not matter to me how you come to that conclusion if we agree the system is indeed overprovisioned.

1 Like

Thank you @crabman for starting this thread!

I have to give this proposal some more detailed consideration, but here is some top-level thoughts. It looks like @RawthiL has already covered much of this.

While I am very sympathetic to Relay Driven Inflation (“demand-centric emissions” to use @vitaly’s terminology from the days where SER was being debated and shaped), I think it is premature to go down that road. Balancing emissions to demand is something that happens in maturity, not during the bootstrapping phase. We are still in the bootstrap phase.

First, true balance between demand and emissions should only consider paid demand. If relays are offered for free to a set of clients, either the gateway providers or the network providers need to absorb that cost. If the gateway providers absorb that cost, then from the network perspective, those relays are paid relays, not free relays and they can be factored into setting emission levels. However, that value right now is zero (PNI does not pay the DAO for relays). If network providers absorb that cost, that means network providers must service those relays without compensation as sort of a “tax” on their paid-for relays, hence those relays cannot factor into setting emission levels.

Even the 200M paid relays that @RawthiL quotes is too high by a factor of two to three because that number only means “some level of payment” not Infura-level rates. Infura-level rates cannot be used for a benchmark. It represents a retail level rate that POKT institutional clients receive from their clients. To secure a client such as Infura, PNI must offer a rate very competitive to Infura’s internal opex, which may be around 0.000004. So lets suppose ~0.000003 to 4 is a realistic gateway charge, gateways have their own opex and need to be profitable, so maturity means being able to offer service at the protocol level for 0.000002 to 3. By the time all is said and done, if we really want to enter into “maturity” phase now, we are talking cutting emissions by a factor of 30x from current levels, not by a factor of 3x. I don’t see the network surviving a 30x slash to emissions long enough to get to the 3B paid demand level. I understand the proposal is not to slash emission to 1/30th current values but to a value approximately 10x higher than this. That is fine, and the proposal can be considered on its own merit, but any pretext of it balancing supply and demand should be dropped.

Second, right-provisioning of the network happens at the granularity of individual regions and chains. Some chain-region combinations in the network are currently over-provisioned, some are right provisioned and some are under-provisioned and rely heavily on crutches such as altruist to obtain reasonable QoS. In v1, there will be capability to balance emissions to demand on a chain/region granularity, but that capability does not exist in v0.

For better or for worse, Pocket Network has staked its reputation and market differentiation on servicing a great multitude of chains. To slash emissions by a factor of 3x from current levels will surely force a further consolidation down to the “most profitable” chain/region combinations. This can’t possibly be good news for chain/regions that currently are right-provisioned or under-provisioned. The reputation hit from degraded QoS that results may be significant and perhaps not recoverable. It seems advisable therefore to wait to conduct the “experiment” of a drastic cut to emissions until v1 at which time smaller chains/regions can be preferentially propped up while focusing the drastic cuts on the high-volume chains/regions.

2 Likes

Qos is already bad, Speed is just one aspect of that.

Please read the proposal on the reasons we need to end this bootstrap phase. We are not really entering a maturity phase with this proposal, just cutting out emissions to sustain us till v1. We will need to potentially bootstrap again if v1 fixes the issues that is preventing the growth of our relays.

Portal currently charges higher than this infura rate on their PAYG model. This is something you have to take up with them and we can revise the benchmark.

1 Like

Thanks. Yes I will look into the proposal more. Also Q1 report will come out in about a week. The numbers will give us a good feel for current snapshot of paid demand. PAYG should not factor much, if at all, into those numbers. I will comment further after a more detailed read.

1 Like