The Avalanche network experienced a slowdown yesterday. I want to provide a short update on what happened, with a full analysis to be provided next week.
The Pangolin launch generated enormous activity, as I’m sure you’ve already heard. This activity triggered a nondeterministic bug in the network layer, related to caches, that led to a block validity check being skipped.
As a result, 57 X-chain to C-chain coin mint transactions were not subject to requisite checking in the client. This led to invalid minting of ~790.2 AVAX, corresponding to ~$40k USD, on the C-chain. It also caused the network to slow down. Nothing else was affected.
This invalid minting was the result of a bug in the client code, not a safety failure of the network or an error in the protocol. It is identical in nature to the Bitcoin underflow bug that enabled the creation of 4 billion BTC and required a patch.
The client was rapidly patched to fix this issue, which enabled all nodes to quickly regain consensus at the same tips.
Code is law. No funds were lost, and because Avalanche is a permissionless and immutable blockchain, no transactions were or could be rolled back.
Avalanche explorers may show erroneous information until patched to handle this unexpected event. Interested parties can verify the foregoing by looking through the fix: docs.avax.network/build/release-….
Overall, we are proud of the speed and nature of the community response. If Avalanche had been centralized, the mint could have been reversed and the bug would have been resolved instantly.
All systems pioneering new technologies will hit road bumps, and Avalanche is stronger today as a result.
And the people who received 790.2 AVAX, enjoy your bug bounty! Love you all for being early adopters ♥️.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Another #FreeLoveFriday. So far, I’ve covered Bitcoin, Mastercoin/Omni, and last week ChainLink and the importance of decentralized oracles. Today, let’s talk about one of the most fascinating projects in crypto - @MakerDAO
In my thread about Mastercoin, I briefly touched on the vital role fiat-backed stablecoins play in crypto markets, but there’s a catch with them:
The counterparty risk of a third-party holding fiat in reserves.
Enter MakerDAO, which set out to create a decentralized, collateral-backed cryptocurrency, DAI, that would be “soft-pegged” to the U.S. Dollar using the power of algorithms. In crypto tradition, its supporters said trust game theory, not operators.
Important reminder about @RobinhoodApp's past track record.
Robinhood has a history of failing to provide "best execution" to their clients. In essence, they let hedge funds like Citadel and others front-run the orders from retail, costing millions to the retail investors.
Are DEXes better? Undoubtedly. But DEXes on slow PoW chains like ETH1 just don't cut it, because they allow miners to front-run the orderflow. hackernoon.com/front-running-…
Back with another #FreeLoveFriday. Last time, we covered how Mastercoin/@Omni_Layer pioneered digital asset issuance on blockchains. Today, let’s discuss @Chainlink and the vital role it plays in connecting blockchains to the real world.
I have said repeatedly that digital asset issuance is the killer application for blockchains. The next frontier is bringing real world assets to networks like @AvalancheAVAX, but we often face a significant problem:
Namely, how do you get data from the real world onto blockchains and into applications running on them? More critically, how do you achieve that securely and transparently in real-time? Smart contracts are tamper-proof, but they're only as reliable as their input data.
In Bitcoin, a transaction isn't final until it's in a block that is k deep. k depends on exchanges and is 3 or more, with 6 being a typical choice. Since the initial block wasn't that deep, a "spend" didn't happen, and therefore there could not have been a double spend.
Now, the choice of k depends on a few factors. 6 isn't a magical number that's correct for all time. It depends on the amounts of hash power available to the attacker. If the attacker has access to 49% of hash power, k should tend towards infinity.
The tail of any PoW blockchain is kind of like a scratch/working area. Changes there are to be expected and perfectly normal.
Remember that PoW's safety depends solely on the amount of hardware that is available to launch a 51% attack. If an attacker has 51% hashpower, the number of confirmations required for safety is infinity -- the coin is not safe to use.
Changing the hash function is something that people try, but it typically doesn't work: once the coin is turned into a GPU-mined coin, the attacker has as much hardware to attack with as there are GPUs to rent.
Back with another #FreeLoveFriday. My first thread focused on what I love about Bitcoin, and features we borrowed for @avalancheavax. Today, let's focus on @Omni_Layer, or as OGs knew it, Mastercoin
Let’s wind back the clock to early 2010s. Bitcoin is just getting started. Deep techies and cryptography people are hearing whispers, reading Satoshi’s whitepaper, and many are getting hooked on the idea that money can be decentralized.
But why stop at money?
In January 2012, J.R. Willett publishes “The Second Bitcoin Whitepaper v0.5”, which laid the foundation for what has always been the absolute killer app for blockchains: digital asset issuance sites.google.com/site/2ndbtcwpa…