Proposal: Increase of 'max_validators' cap

Since the upgrade to akashnet-2 in February 2021, software soft upgrade in early April and enablement of IBC transfers at the end of April, the Akash Network has shown impressive stability and performance in the operation of active deployments.

In the search of continued decentralisation, and due to the large number of competent validators currently sitting outside the active set or waiting to launch their validator, Bliss Dynamics believes it is time to start the increase of the maximum number of validators allowed in the network.

As described in the paper “AKT: Akash Network Token & Mining Economics” released on 31st of January 2020 by Greg and Boz in:
Section ‘II. NETWORK OVERVIEW’
->Subsection ‘A. Proof of Stake Based Consensus’
–>Clause ‘1. Limits on Number of Validators’
We see that “we can support enough validators to make for a robust globally distributed blockchain with very fast transaction confirmation times, and, as bandwidth, storage, and parallel compute capacity increases, we will be able to support more validators in the future.” And also that “the number of validators at time t year will be: Vn(t) = [log2 (2t) · Vi,0]” where Vi,0 = 64 Validators which is also graphed in Figure 2 (page 4):

For those not so familiar with logarithmic functions, the formula suggests a maximum validator cap as follows:
Network Launch (25/09/2020), t0: negative (assume 64 validator starting set)
Today (31/05/2021), t0.68: 28 validators (assume 64 validator starting set)
1 year (25/09/2021), t1: 64 validators
1.5 years (26/03/2022), t1.5: 101 validators
2 years (25/9/2022), t2: 128 validators

As the formula does not well describe the early stages of time, t, and to avoid a large step change and having to increase the number of validators by nearly 60% in a 6 month period (from 1 year to 18 months), we believe it would be worthwhile for network stability to increase the max_validators parameter in the network from 64 to 75.

Bliss Dynamics have reached out to a number of community members and validators both inside and outside of the active set, and believe there is not only sufficient support for this proposal, but also a sufficient quantity of capable validators to immediately fill the slots.

Now we wanted to open the topic up to the wider community for comments and discussion prior to submitting the proposal on-chain. Please let us know your thoughts in the next 2 weeks.

Note: After roughly 2 weeks of discussion, then 2 weeks for the proposal voting period, the parameter change is expected to take place around the beginning of July.

8 Likes

As a beginner in Cosmos, I will study and learn the details,
Thank You :slight_smile:

1 Like

Further background:
Cosmos Hub
Cosmos Hub proposal 10 ending voting 17 July 2019 stated that:
“In the Cosmos whitepaper, it states that the number of validators on the Hub will increase at a rate of 13% a year until it hits a cap of 300 validators. We propose scrapping this mechanism and instead increasing the max validators to 125 validators in the next chain upgrade with no further planned increases. Future increases to the validator set size will be originated through governance.”

The proposal passed with a 63% turnout and 95% yes vote. To this day, the maximum number of validators cap on the hub remains at 125.

While what works for another chain may not necessarily be suitable for Akash, it’s still helpful to understand what others have gone through before.

Other Cosmos SDK chains’ max_validator cap
IRISnet: 100 (Fully occupied, proposal to increase from 100 to 108 in April 2021 did not reach quorum, 49.49% turnout compared with threshold of 50%, although votes were 99.98% yes)
Terra: 130 (Currently 108 active, proposal to increase from 100 to 130 in May 2021 [4 days ago] passed with 51.9% turnout and 91.3% yes vote)
CertiK: 125 (Currently 94 active)
Oasis Network: 101? (Fully occupied)
Crypto.org: 100 (Fully occupied)
Kava: 100 (Currently 88 active)
Band: 100 (Currently 60 active)
Sentinel: 64 (Fully occupied)
e-Money: 50 (Fully occupied)
Regen Network: 50 (Fully occupied)
Secret Network: 50 (Currently 48 active)
Starname: 16 (Fully occupied)

Based on our initial analysis and conversations, we believe that with a raise of the max_validator cap to 75, Akash Network will still be fully occupied or nearly fully occupied almost immediately following the parameter change.

StakeLab plan to join Akash Network if the proposal is validated.
We will be up for at the time of the upgrade.

If some validator plans to join too thats the moment !
Maybe Akash team could help get more vision on this with the form to apply for next validators to onboard ?

1 Like

EZStaking also plan to join Akash Network if the proposal is validated (already ready and in the inactive list).

1 Like

Hello,
Dual Stacking currently is join, but not in active set. I vote for this proposal.

1 Like

We wanted to raise and address 2 potential reasons why someone might prefer that this proposal be changed or postponed prior to it going live on chain (neither of which we believe are supported by the community from our discussions so far by the way) but happy to hear more input here on the forum:

  1. Timing around so much else going on:
    While launch on Osmosis is imminent, and teams’ activities in Miami have just finished, it’s certainly a busy time for the Akash community. As there is no network upgrade or significant effort required from validators (except to vote) as the parameter is automatically adjusted following the result of the proposal, we don’t believe this proposal puts any undue pressure on any of the community at this time and should go ahead on the schedule as originally planned.

  2. Change the cap from 64 to 100 validators instead of 75.
    Some chains have elected to increase their cap significantly and allow the validator community to grow organically into the cap over time. Increasing the set to 100 (by over 56%) and hence providing empty validator slots could be bad for the safety of delegators (it becomes too easy for amateur or nefarious validators to launch on Akash without skin in the game or a network of support) and since a further increase of the set is only another proposal away, we believe 75 is the appropriate cap size at this time.

Also, although we have received some support already privately (thanks a lot!), it would be great to hear from some more existing validators here regarding how they plan to vote on this proposal.

Next steps: We intend to create a draft proposal wording later this week (it will be simple and straightforward) in line with the original timeframe which would see the proposal being launched on chain around this weekend.

Note: Bliss Dynamics plan to put up a majority of the deposit and then give some time for validators to review the proposal and support with their own deposit if they choose.

I support this idea, 75 set seems to be fine, although 98 could be better.

3 Likes

Hi Greg,
Thanks for the support. Interested to hear why you like a cap of 98?
While I think we could be almost immediately full if we rise to 75, a larger set would allow us to attract more validators over time without having to go through another proposal.
And would 100 be a more round number to make us engineers feel better? :sweat_smile:
Open to suggestions and more comments.

2 Likes

I like the idea of increasing from 64 to 75. The reason for this is we’d like a gradual increase over time, while ensuring the implementation of proactive delegation framework that contribute to network growth.

64 to 100, imo, is too aggressive for a young network like ours, and for it to be effectively managed, especially with regards to foundation delegations to help bootstrapping validators.

We’re aiming for quality of validators over quantity, and a gradual increase, over time, would be ideal.

3 Likes

My goal is to become a validator of the network, so the more spots the better! :smiley:

1 Like

Hi, yes I think it is a great idea to increase it to 75. As Boz said, quality over quantity.

4 Likes

Hello, Jetlife Staking support the proposition for 75 validators.

3 Likes

seems it’s reasonable.

2 Likes

Agreed with 75 validators. For current validator list, only one inactive validator with high enough stakes though

1 Like

Hi Bitoven,

The only cap is on number of validators, not any specific number of AKT delegated. This means that if the 75th validator on the list by delegation has only 100AKT if the proposal passes, then this is enough to get them into the active set (assuming their nodes are online).

As you note, one validator has over 300k AKT delegated and is outside the active set, but all validators down to 75 will also be automatically included.

Thanks everyone for the comments.

2 Likes

It sounds great to me, I’ve been an Akash investor for several months and already validating for other Cosmos networks and planning to join more, Akash would be my first choice to go.

For me 98 is the best.

There are already 98 validators (active + inactive) and all the new validators that would want to participate, so the max cap would be reached for sure.

About the staked amount of the inactive validators, you cannot expect all the new validators to be rich and have a huge stake without the opportunity of being delegated. To me blockchain is about decentralization.

1 Like

I will support the proposal as I believe it brings further decentralization to the network. It is a great opportunity to welcome more motivated validators, who will contribute to the growth of the ecosystem!

4 Likes

Support to increase validators from 64 to 75, quality over quantity!

2 Likes

I agree in principle that we can certainly increase to 98, but we’d prefer 75 for an extremely young network.

We can get to it, but let’s increase it to 75 as a first step together, learn as we grow, and iterate on improvements for coordinating 75 independent teams to map out processes and communication plans etc., to make it seamless. From 75, we can re-evaluate at a later date to further expand on the cap.

We’re also going to be releasing and sharing a delegation framework to ensure foundation delegations are used effectively versus passive-validators that do not engage with the community or network in a meaningful way.

4 Likes