1/15 Avalanche Consensus uses a DAG data structure, runs on the X-chain, a no Smart Contract network with a simple "Account to Account" Token Transfer VM. This Consensus Mechanism is unable to do a total ordering of transactions. So is it really Distributed Consensus? 🤔
2/15 Avalanche figures out that its Consensus is unable to totally order transactions, a requirement for Smart Contracts. They spin out a SC C-chain, throw the DAG away, and create Snowman Consensus on a linear #blockchain.
Easy right? One block over another. Order solved 😊
3/15 Oh but wait! What if 600 nodes are gossiping at the same time. Which block should we append at height X?
That's easy too! Throw the Leaderless part in Avalanche, upgrade to Snowman ++ so you wouldn't need to do extra work to achieve consensus as Prof. Sirer confirms below.
4/15 "Snowman++ is really a major change in kinda how we propose blocks, and it's kinda one of the things we've been working on with an eye towards improving throughput, latency and keeping the requirements low for validators from a hardware side" 7:30
5/15 "Anyone could propose a block for any height, and if we do, then the system has to choose between the 600 of us. But with Snowman++, we all say look, for this height, we're going to have Steven take the lead and if Steven proposes a block, we have a preference for it" 15:25
6/15 @ercwl knows Avalanche has "no slashing", finds no issues with that.
"Many systems out there that don't have slashing...And yet they work just fine" 37:10
"There already exist chains, especially the one that you're trying to copy already doesn't have slashing in it!"
37:40
7/15 "As powerful as the DAG is, the core limitation of it, is that you can either provide a Partial Ordering...
So we ended up implementing Snowman as a 2nd consensus (@stephenbuttolph the DAG has no limitation, Hashgraph totally orders it just fine) 7:40
8/15 What if I had to transfer assets from P-Chain, X-chain to C-chain for SC. Listen in video above at 28:40 to 29:35.
That's what happens when you have a half-baked Leaderless DAG with a "Metastable" Consensus.
You start off with one chain and end up mitigating with 3 chains 😂
9/15 Just when you think Snowman++ was done, you discover Avalabs waving #Frosty at you, you'd have to wonder how Meta"unstable" Avalanche is, so they keep trying to improve it. So do we have Consensus over what our Consensus should be or not yet? 😅 forum.avax.network/t/frosty-conse…
10/15 Hashgraph stands tall, an unwavering invention from the mind of Dr. @leemonbaird. An elegant and efficient algorithm with a Math Proof for its aBFT properties.
Behold the difference between the crippled #Avalanche DAG and the #Hashgraph DAG!
11/15 Hashgraph is so amazing, even stuttering @el33th4xor admits "#Hedera#Hashgraph is a fine Protocol" 56:10
Unfortunately he gets everything else wrong, or intentionally misleads his audience claiming Hedera is a rehashing of 1999 pBFT protocols. 39:18
12/15 The truth is Avalanche is more PBFT like because it queries the network through repeated rounds and sub-sampled VOTING.
Hedera uses proven aBFT algos to achieve consensus without voting. Nodes VIRTUALLY VOTE on behalf of other nodes locally.
Get your facts straight Prof.
13/15 Hashgraph is not "strictly" slower than pBFT. pBFT protocols may achieve a 1000 tps, but rightfully cannot scale the node set. Hedera uses no vote, can scale to 100s of nodes within 1 Shard, with 10,000+ tps throttled, and Nodes dynamically joining the network, at any time.
14/15 Seems Avalanche wanted to hop on a trend 😂
"There's generally a trend towards protocols that build graph like structures because graphs can be built much faster, because I can extend 1 edge of a graph while other people are extending other edges of the same graph" 57:15
15/15 Describing a DAG forking is surely trendy, yet Consensus was nowhere to be found! 😅
No wonder AvaLabs had to upgrade Avalanche DAG for Blockchain Snowman (Whitepaper?), a Leader-Based PBFT like consensus.
SORRY! But I'm hodlin my $hbar
Happy New Year #Hbarbarians!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
2/5 "One of them, a member of the @txblockchain1 would be #Hedera#Hashgraph, they are not a #blockchain in the traditional sense. They don't actually use the blockchain structure we talked about, they use a different structure, the #hashgraph..." 1:33:50
3/5 "They actually have what's called #HCS, you can almost think of it like a distributed Kafka Cluster...You can essentially use this for your private blockchain, whether it's HF, Quorum or Corda...Another interesting approach for a hybrid solution" 53:13
1/7 Nephrite Jade is very rare and its collection value price is more than US$15,000 per Kg. JDX cooperated with Dr. Li Dezhong, founder of JPin, to collect several pieces of Emperor Wu's Hetian Jade of the 2000 year old Han dynasty. #Hedera is a partner collaborator
2/7 @peterkuoGOP founder of JPin is an investor with a focus on blockchain & an extensive experience in tech start-ups. He is also Vice Chairman of CA Republican Party, GM of Silicon Valley VC Roundtable, President of HRSV Rotary Club, GM of NAIFA
3/7 Curation and IP Licensing:
JPin will hold a global exhibition with high quality antique jade to drive up the value of the collection, which will be funded through authorizations of IP addresses and the sale of peripheral products based on the exhibits.