By Dionysis Zindros 🐼
The Harmony Research DAO is the research arm of the Harmony ecosystem. Following the vision of Harmony towards openness, transparency, and decentralization, Harmony funds the Research DAO to fulfill its research objectives. As a Decentralized Autonomous Organization, the Harmony Research DAO is independent and takes decisions based on its own mission. In addition to the initial funds provided by the Harmony treasury, the DAO is also funded by its own members to help steer the research directions that will benefit the community.
The mission of the Harmony Research DAO is to promote and fund theoretical and applied peer-reviewed research of excellent quality in the field of blockchain science. Our vision is to establish the field of blockchain science as an independent and recognized field in Computer Science, spanning across the areas of cryptography, distributed computing, security, programming languages, networking, and information theory. We are a non-profit DAO whose goal is the betterment of the field as a whole. We are recruiting researchers from a multitude of fields to pursue theoretical and applied research that will have a positive impact on the Harmony blockchain itself, but also the broader blockchain ecosystem. Since Harmony is about creating an interoperability network between all chains, any work that helps improve the broader blockchain community is welcome.
The primary means by which the DAO promotes its goals is by funding:
- The development of research papers by researchers in the field.
- The conduction of experiments required for taking measurements and collecting empirical data.
- The implementation of research directions in code to ascertain their viability.
- The organization of and participation in conferences in the field of blockchain science.
The DAO also promotes its goals by keeping abreast with latest developments in research on blockchain science and steering the direction of the field by making announcements and running contests in which research priorities are established.
Central to the decision taking of the DAO stand its core values:
- Transparency. Funding decisions, logistics, and accounting data are published openly. Anyone can review the capital allocation and the endowment of the DAO. All funded papers, conferences, and projects are public information.
- Openness. We mandate that the research funded by the DAO is free of software patents and published under a Creative Commons licence. Any software developed in collaboration with the DAO is licensed under an open licence (GPL, MIT, BSD, or Apache). We also mandate that any papers funded by the DAO are published in open archives such as ePrint and arXiv. In this manner, the Harmony ecosystem, but also the broader ecosystem can make use of our research results.
- Rigour. We value research that abides to academic rigour. We prioritize papers that are peer-reviewed and conferences that adhere to high academic standards of peer reviewing integrity, following the double-blind process of peer review. The two primary peer-reviewed conferences in the blockchain field are currently Financial Cryptography (FC) and the ACM Advances in Financial Technologies (AFT). We are also recruiting works in the top security and cryptography conferences in the field.
- Impact. We care about publishing high-quality, impactful papers that solve foundational and difficult problems in the field. We want our research to help the field move forward. If a problem is central in the field, has been identified by the wider research community, and remains unsolved, we want to tackle it. We are especially interested in problems that impact multiple different blockchain systems and the field at large.
- Equality. We fund research that follows the principles of the cypherpunk community. Research that increases the power and well-being of everyday people, supports the less well-off among us, respects their privacy, enables the sharing and freedom of information and content, and increases the accountability of organizations and governments.
The Research DAO stands at the intersection of three fields of computer science:
- Cryptography. Cryptography concerns the design and analysis of protocols in the presence of adversaries. A strongly mathematical field, it touches upon both low-level primitives such as signatures, encryption schemes and zero-knowledge proofs, as well as higher level protocols such as blockchains, consensus algorithms, and authenticated data structures. Prominent academic conferences in the field are CRYPTO and EUROCRYPT.
- Security. Security is the applied field that ensures systems are protected from attackers. Taking cryptography and deploying it to real-world infrastructure, secure systems are resilient to attack and verify this empirically through penetration testing. They also check this formally using formal methods, and ensure this in practice with hardened programming languages, protocols, and APIs. Prominent academic conferences in the field are ACM CCS, IEEE S&P, USENIX Security, and NDSS. Community events are the CCC and DEFCON. An industrial conference is Black Hat.
- Distributed Computing. In the community sometimes referred to as simply “decentralization”, this field is about designing protocols in which multiple parties coordinate to achieve a common goal such as reaching consensus, without trusting a central third party. Throughout its history, the field has explored Byzantine Agreement and other basic consensus algorithms. Today, it has been revived with the invention of the blockchain. Prominent academic conferences in the field are ACM PODC and DISC.
Blockchain Science stands between all three: we are developing new cryptography to make blockchain systems possible; we compose cryptographic primitives to build high-level secure systems; and we build them such that multiple parties can coordinate to achieve a common goal without a trusted third party. Naturally, other fields of computer science and mathematics come into play also, including privacy (with landmark conferences such as PETS), information theory, programming languages, networks, and game theory.
As a new field of computer science, Blockchain Science has many interesting and foundational outstanding problems. As a Research DAO, we have identified and will be pursuing solutions to the following areas:
- Bootstrapping. This concerns the speed and efficiency with which wallets synchronize with the rest of the network. There has been significant advances in the last few years on the topic of fast synchronization in proof-of-stake blockchains, including the works on Non-Interactive Proofs of Proof-of-Work (NIPoPoWs) using superblocks, FlyClient, and logarithmic space mining. Open problems concern the security of NIPoPoWs in the variable difficulty setting, but also the development of proof of proof-of-stake (PoPoS). A topic with significant theoretical depth as well as many practical applications, tackling this problem can yield exponential improvements in how long it takes to synchronize a mobile client. At the same time, it allows removing centrally trusted servers without harm to performance. Developing proper “superlight clients” also allows for building trustless cross-chain bridges, without the need for trusted federations and overcollateralization.
- Interoperability. The number of chains, coins, and protocols keeps growing and growing. Ensuring that these can play well with one another has become a central problem of the space. It has become clear that there will be no one coin to rule them all, but multiple coins working in tandem, each offering its own unique features. Developing an ecosystem of collaboration, in which different protocols can speak to each other and interact in a secure and performant manner is both a scientific and an engineering challenge. Between main chains, communication must be done both between proof-of-work and proof-of-stake chains. Making use of bootstrapping techniques to build cross-chain clients, moving data from Layer 1 to Layer 2 and back quickly and securely, as well as communicating between the real world and the chain world using oracles are central questions that fall under this topic.
- On-chain scaling. The main chain functions as the settlement layer and all parties reference it for finality. Scaling this layer has become the current main challenge of our science. There are several means by which scaling can be achieved. With sharding, a blockchain is split into multiple subsystems, each with its own validators. Ensuring the validators are allocated in a secure manner even against an adaptive adversary is difficult. Another means is developing authenticated data structures that go beyond the concept of simple chains. From parallel chains that cross-reference each other, to DAG-based systems, such topologically exotic consensus systems are full of promise.
- Off-chain scaling. Scaling a blockchain’s Layer 1 infrastructure can only get us so far. For achieving the desired scalability of a global monetary and contract system, most transactions will have to be moved off the main chain. There are many candidate approaches here. Sidechains and interoperability between them would allow the creation of smaller chains that can take off some of the load. Payment and state channels can allow the transaction of smaller groups of people off the chain, but also develop more globally as payment and channel networks are built on top. Lastly, rollups of the optimistic and ZK kind have seen significant adoption in the last year, and are prominent candidates for scaling data off the chain.
- Consensus. The foundation of chain protocols is always an appropriate consensus mechanism. We have consensus protocols employing proof-of-work and proof-of-stake, and many have been proven to be secure. Can these protocols be optimized from first principles to achieve better performance without harming security? Questions such as increasing block sizes and block production rate, or changing the longest chain rule, temporary dishonest majority, as well as applying concepts from information theory pertain to this area. This area also pertains to the analysis and consolidation of existing consensus protocols, from the era of byzantine agreement to today’s complex decentralized consensus protocols. Lastly, it also pertains to the development of theoretical tools to aid the understanding and education around consensus.
- Formal verification. Developing new protocols is only one aspect of ensuring they are secure. In addition to the mathematical tools in the arsenal of cryptography, tools from the area of formal verification can be used to ensure that both mathematical proofs are correct, through the use of a proof checker, but also that the software implementations of such protocols really does follow the protocol as intended. Very closely tied to these concepts is the development of secure programming languages for smart contracts that lend themselves to such tools.
- DeFi. The smart contract ecosystem is evolving to replicate all of traditional finance, and beyond. Many concepts that are already possible in Decentralized Finance (DeFi) are new and have never appeared before in traditional finance – such as flash loans and perpetuals. Other financial derivatives such as options and futures as well as useful instruments such as insurance, payroll and loans are being developed too. This very new field raises a plethora of open questions in security, from contract composition, to oracles and miner extractable value. Another question pertains to the fair governance of all these protocols – starting with our own DAO. Lastly, the proper deployment and upgrade of these contracts and the underlying blockchain to support new versions remains a central problem.
- Networking. Blockchain consensus typically models the network as a simplistic machine. However, the devil is in the details. Burning questions, such as achieving order fairness, with the involvement or not of a central trusted party, are becoming more and more important. Tradeoffs between performance and security, reducing latency, and taking full advantage of the available bandwidth are central here. As everywhere else, a powerful adversary may also be able to disrupt the network, and protections against splits remain a central question. But given a more lax adversary, better efficiency may be possible. Electing temporary leading parties can also help.
- Economics. Consensus protocols can work under honest assumptions, but how are they incentivized? In particular, is each participant’s financial gain aligned with the consensus protocol’s properties and goals? This touches on the field of game theory, with many questions remaining open. The topics pertain to pool formation, delegation of participating rights, sybil resilience and resilience against malicious coalitions. In the midst of all this comes the topic of building and governing a transparent macroeconomic policy, upgrading it, and controlling money supply without a central bank. Lastly, difficult questions, such as egalitarianism and fair allocation of rewards, arise also, some of them with more philosophical and ethical ramifications than we initially imagined.
- Privacy. Blockchains are the first practical application of zero-knowledge proofs, a much loved if not idolized concept in cryptography. The ability to perform private transactions that enable untraceability and unlinkability is one aspect. The ability to have fully private smart contracts and smart contract state, whether on layer 1 or on layer 2, is a much more difficult goal. New blockchain-centric primitives that enable zero-knowledge creation of stake and signatures are central to these systems.
- Usable security. Even if we build the perfect systems technically, in the end our users are humans. The current state of affairs in blockchain systems is disconcerning: Most truly decentralized wallets and other end-user software are barely usable. To make matters worse, the inherent irreversibility of blockchain systems all but ensures that small mistakes might have devastating effects on the users’ accounts. The topic of usable security concerns the human/computer interaction in blockchain systems, helping users understand what is going on at every point in time. Having easy to use wallets, social wallets that cannot be easily lost or stolen, multi-factor authentication, reasonable spending limits, and easy hardware wallets are key questions here.
- Community. While systems are designed with decentralization in mind, this is often not achieved in practice. To ensure proper decentralization, concrete metrics must be proposed and relevant measurements must be taken and experiments conducted. Usage metrics on staking, mining, network and node decentralization allow us to collect such statistics to gauge whether decentralization has been achieved and, if not, seek ways to rectify.
- Transparency. New institutions are replacing old ones as decentralized finance is taking shape. Blockchain systems and DAOs must be governed for the people by the people. Centralized organizations such as exchanges must also be held accountable, while maintaining privacy. To ensure these, regulatory transparency must be ensured by developing privacy-preserving proofs of assets, liabilities and solvency, as well as off-chain transaction auditing. The tools to do that securely and privately are an important research topic.
These 13 areas of research in Blockchain Science will be central for the next few years. Harmony’s Research DAO will fund, support, and steer the direction, so that the foundational problems in all of these areas are tackled with academic rigour and an eye for application.
From Harmony's DAO Guidelines:
Governors of each DAO have delegated autonomy over its assets and initiatives. Harmony helps define 3 broad mandates, recruit 9 governors, define the deliverables and metrics for the first 3 months, and fund at maximum $1M. We recommend $75 to $350 per hour as the self-assessed salary, 3-month election terms, retroactive peer bonus and performance feedback, 80% passing votes, and openly tracking timesheets and deliverables for each member.
In 2021 Q4 Dionysis Zindros will lead this Research DAO. After 9 governors are appointed, Harmony will send $100K initial funding (leading to $1M in total by 2022) to a Gnosis Safe's multisig address held by the governors.
Mandates – Bring Best (Research) Results to Production
- Cryptography: deploy succinct proofs for privacy & performance
- Security: use mechanical verification for audits, strongly typed languages for prototyping
- Decentralization: scale transactions across protocols, on-chain staking & delegation across light clients
Governors & Deliverables
- Dionysis Zindros (at decrypto) – toward logarithmic states: Flyclient/NiPoSPoS-based bridges vs ZK-rollups vs Interlay/XClaim's economics security, fast state sync for mining and resharding, keyless wallets with lattice-based witness encryption;
- Dimitris Karakostas – toward supporting 100k delegators: optimizations of Harmony's on-chain delegations and compounding rewards;
- Aaron Li – toward end-to-end formal verification: prove security of Harmony's authenticator-based wallets in Coq, smart contract audits on Horizon bridges for Ethereum and Bitcoin;
- Zeta Avarikioti – toward 1-second transaction finality: determine the optimal shard size against network security for our Proof-of-Stake network, compare to Mahdi Zamani's Rapidchain and Instachain;
- Ivan Homoliak & Andrianna Polydouri – toward authenticator-based wallets: client-side encryption security and performance, one-time-password (OTP) Merkle-tree generation and authenticator security.