“Harmony” Review and Rating

Harmony Technical Whitepaper (v.2.1) produces on me a positive impression. In short, Harmony aims to upgrade Zilliqa’s sharding by PoS, segmented storage data and improved randomness generation. There are more to Harmony, than that, but this brief review format doesn’t allow to dig too dip into this 21 pages work.

This paper is, definitely, well-written and structured. Moreover, its authors not only managed, in a relatively limited space, to comprehensively present their new consensus protocol (which they called “Fast Byzantine Fault Tolerance (FBFT)” and which, basically, instigates so-named “leaders” to fast-track validation by collecting validators’ votes into the multi-sign and then broadcasting it) but also to critically analyze FBFT and to compare it with already existing protocols. That is a rare occasion in our generally very sketchy and confused “papers writing” world and I can’t under-appreciate that fact.

Now lets see how it reflects on the Harmony’s Security-Velocity- Engineering-Transparency sub-rating.

As we all know there’s not such thing as the absolute “Security” and even systems with a “proven” and long term track record, time after time, had showed themselves to be vulnerable to the new types of attacks. Because of that assigning “A” rating to the new protocols would be, IMHO, too generous.

Nonetheless, in this piece authors argue that their proposed amalgams, called “Scalable Randomness Generation with VRF and VDF” might improve upon Ethereum 2.0 design with VDF (Verifiable Delay Function), Algorand’s one, which employs VRF-based (Verifiable Random Function), and Omniledger, which uses the RandHound protocol. As I said, I don’t think it’s possible to fully substantiate “security” on paper and I leave it to you to get yourself convinced (or not) by reading the original text. However, speculatively, my current rating for Harmony’s “Security” is “b+”.

The “Velocity” part of sub-rating intends to asses system’s scalability and throughput. Harmony, basically, tries to boost the later by adding blockchain state shards on top of network traffic’s and transactions’ ones. Additionally, Harmony is based on PoS not PoW, which also makes reaching consensus much faster. Besides, fostering scalability, Harmony introduces cross-shard transactions. We have yet to see how it’ll work on practice under different traffic regimes but I’m tempted to assign “A — “ to Harmony for those good and, IMHO, well-thought-trough efforts.

The “Engineering” is a sub-rating parameter, which is more “aesthetic” that “technical”. It intends to reflect how new, elegant and innovative the proposed systems and protocols architecture is, or, better say, seems to be.

As authors put it themselves: “Harmony addresses the problems of existing blockchains by combining the best research results and engineering practice in an optimally tuned system.” That’s exactly how I feel myself about it. When I read this paper I was very pleased by its professionalism and clarity. However, I was not dumbstruck by its innovativeness or novelty. That makes for very solid and respectful “b” on this side of the sub-rating.

The fourth part of the sub-rating “Transparency” aims to measure the degree of centralization / decentralization of the proposed system architect. It presumes that the high level of centralization reduces transparency preventing wider adaption of the system. This parameter is quite hypothetical and I intend to elaborate on it in the future.

However, now, I feel that this post has already became overextended and I will just say that Harmony “Transparency” is “b+”, not “a”, because, I don’t think that even the proposed solid system architect, based on PoS, improved sharding and DRG is fully protected against future “centralization” attempts entertained by the large stake holders.

Security-Velocity-Engineering-Transparency (s.v.e.t.) sub-rating: [b+a — bb+]

If you want to know more join whitepaper review and analysis group: https://t.me/joinchat/I5eQ-A6FSC2vXg_PNgFwJw