Chrysalis Phase 2: Transactions Size Will Decrease By 20 Times

The Chrysalis Phase 2 has been unrolled and now the transaction size within the Tangle will decrease from 5 to 8 kilobytes for a bundle to 350 bytes for the atomic transaction as we are reading more in the IOTA news.

The messages per second will also increase by about 20 times. The IOTA developer Hans Moog revealed new details about Chrysalis Phase 2 via Discord. As it was announced, the upgrade involves a change from the bundle concept to atomic ones. The founder David Sonstebo stated in February that the upgrade will be accompanied by numerous improvements:

“This will reduce network overhead, reduce signature verification burden, improve spam protection and congestion control, and reduce the length of Merkle proofs (for future sharding). In addition, it will reduce implementation efforts and increase the maintainability of our core node software.”

Until now, IOTA used the concept of bundles to create transactions as they are transfer packages that contain one input and one output transaction which can include zero value transactions. The individual transactions were linked to each other via the “trunk.” Bundles are atomic transactions which means that either all of the transactions are accepted by the network or none of them is. At least 3 transactions have to be used to create a simple transfer, 2 for input and the signature, and one transaction for the rest. The problem with the concept is that these bundles have a fixed size.

The concept will be replaced with the Atomic transactions with Chrysalis Phase 2 so that each transaction is atomic. As Moog explained, the size will be more than 20 times smaller. As a result, messages per second will increase by 20 times:

“But I agree bundles are bullshit and that’s why they are removed next – currently the bottleneck is network throughput and ternary conversion.”

Moog also said that the IOTA foundation has already started working on the specifications for the new consensus model dubbed “tangle Multiverse:”

“It’s true. We are writing the specs now. […] Well I have been thinking about it for a really long time but never had the time to write any specs as we were supposed to work on fpc. So it’s very few open questions about how it should work. Rather than ‘does it work’.”