209 похожих чатов

Would you tokenize $million assets using a protocol that isn't

aBFT?

https://twitter.com/AlgoFoundation/status/1443134560665997315?s=19

21 ответов

23 просмотра

I thought TOKO used Hedera, given Council member status etc.

Coinman the HBARbarian- Автор вопроса

It does. Perhaps they have clients asking to tokenize on Algorand?

Coinman the HBARbarian- Автор вопроса

"Algorand network not only complements TOKO’s vision to be an omni-chain issuer, but also brings access to new global infrastructure partners representing a diverse offering of asset classes."

I think Algorand is aBFT. It’s only named in a different way, if I remember correctly. “There is space for others in the room”

This is awesome! TOKO uses HCS so no matter where you want to start Hedera will see the benefits of this usage. Very pleased to see this announcement.

Algorand is a clear competitor to Hedera and offers something closer to "decentralization" but that brings other vulnerabilities... But, also, there is a useful marketing value for TOKO here by attaching to another popular crypto.

I'm guessing that the implication here is that TOKO will mint NFTs on Algorand as well as Hedera. We should understand that council members are not exclusively bound to Hedera. But I'm guessing this will drive business to Hedera in the end

If you use TOKO you will be using HCS so no matter the asset your audit trail is still using Hedera. Big world out there glad to see the utility of Hedera will be used for more and more things!

https://www.reddit.com/r/AlgorandOfficial/comments/nrat1u/algorand_vs_hashgraph/

Thanks for posting this. This is an excellent well-informed comparison. And I think (from my non-technical understanding) that they are basically right that Algorand's resilience claims seem to be equivalent to aBFT. The only thing it gets wrong is that Hedera (like Algorand) achieves consensus based on 2/3 stake not 2/3 nodes. So for Hedera, less than 2/3 nodes could achieve consensus so long as they control over 2/3 stake... Algorand seems to be an excellent platform, and the only one with performance in Hedera's league... But Hedera has seen more extensive testing and, for now, its permissioned nodes hosted by large independent companies offers more security and guarantee of liveness

great article. so it seems as the node count increases the performance of HBAR starts to decrease unlike ALGO

This is true. Hedera latency gets longer (slower) with more nodes. Algorand just randomly chooses a minority of nodes to participate in each round, so performance is not affected by total number of nodes.

Hedera addressed Latency and node count in the last town hall. The answer Leemon gave was that Hedera is able to reduce the time to finality. He noted that it had recently increased above 6 sec and they brought it back down, something which they will be able to do later, when anyone can run an anynmous node. The reason Hedera doesn't deal with the fixes now to ensure 3-5 sec finality now is because nodes are currently permissioned and the team have other improvements they are working on. When they need to reign in time to finality they are able to tweak the code and do so

do you remember how they reduced the latency with more nodes?

cool 😎

The issue I have is that I don't know who the author is or if they are credible, but thanks. I am not going to base anything off of a reddit post but I will dig deeper.

I will dig into this. aBFT is aBFT and anyone in the industry knows that you use that term so making up another term and calling it equivalent is suspect to me.

The actual paper by Wang on Algorand https://arxiv.org/abs/1905.04463

the abstract is good enough to assume ALGO is not so strong.. there was a paper also on attacks against hedera. I don’t remember what was the conclusion however

Yes I posted the paper on attacks on Hedera (search Vienna) a couple days ago. I'll comment on it more soon

Thanks

Похожие вопросы

Обсуждают сегодня

Карта сайта