Notional IBC Over 54% of total successes

Conversation Validation
1

Hi I personally relay 54% of successful IBC transactions.

Here is my validator:

https://www.mintscan.io/cosmos/validators/cosmosvaloper1083svrca4t350mphfv9x45wq9asrs60cdmrflj

I made a mistake setting its comission, Notional’s comission is 9.69% on all networks. I’ll get that fixed in a day or two. We’re more expensive because we’re better.

ICF, other large delegators, looking to delegate to improve quality on cosmos, where are ya?

Here is proof of my claims:

git clone https://github.com/faddat/notional
cd notional
tar xvf proof.tar.gz
grep -o -i success proof | wc -l
grep -o -i already proof | wc -l

Most recently, when I ran this I got:

Screen Shot 2021-08-12 at 11.55.43 AM

Or 54.1%

BTW, since I consider being the main relayer for many major cosmos networks a problem, I open sourced my entire relaying setup. If you followed the directions above you now know how to run a highly performant relayer, if you were to just read through the documents.

Will you?

What gives y’all?

I relay more IBC than 100% of the 125 gaia validators.

Yes, there are 125 validators on Gaia.

Yes, I do more relaying for Gaia than all of them.

3 Likes
2

Hey jacobgadikian, thanks for the continued support you provide to the Hub and the community.

Looks like your logs are over the course of 10 days and look reasonable to me. Do you have all of these chain accounts funded and all running on the same host?

In any case, I would’ve expected and assumed most major validators to be running their own IBC relayer infra, but without any current incentives (coming soon), this is not too surprising I guess.

At the very least I think the ICF and perhaps IG should be running dedicated and funded relayers.

3
  1. They’re all on a hetzner AX61-nvme
  2. Before that they were on bare metal in Hanoi as outlined here:

https://whimsical.com/validatron-PbUypC8tVMU8DxCFNLdDFu

4

I’d be happy to run relayers for ICF / IG but I am not sure that’s what you mean here.

As you’ll note one of the things I’ve done is produce documentation on how to run a highly performant relayer, and how to run validators at the edge of the network. I don’t think it’s a solution for the “mothership orgs” to pick up relaying.

I think that both validation and relaying need to be made easier for new entrants through a combination of documentation and improvements to tech.

I’ve been criticized as saying “it’s just me running IBC”.

No no, that’s not the case at all. Just, I think I am running a dangerously high share of IBC, to be sure.

So basically, IMO, we want many highly performant relayers, and a widely distributed understanding of how to run them.

We don’t want the motherships taking charge of the ecosystem, IMO to be meaningful, this needs to come from the edges.

NB: This isn’t based on a misunderstanding of the first past the post nature of IBC. All this is based on observable declines in service quality that have happened on two occasions:

  • When I changed my relayer address from

cosmos1083svrca4t350mphfv9x45wq9asrs60cg0hunp

to

cosmos1yeac5tgm4mqwl4fyrqp34s0gq5fy8ugxcj76vc

  • A day ago when I ran out of atoms in cosmos1yeac5tgm4mqwl4fyrqp34s0gq5fy8ugxcj76vc

Now, I’m not saying no one else is working on relaying. I am saying that one relayer experiencing downtime shouldn’t harm the whole ibc network, and it clearly does.

2 Likes
5

49.1% now

Chandra Station has built a 0ms relayer and is now operating it.

2 Likes