Algorand’s Forthcoming Technology

Silvio Micali
Algorand
Published in
9 min readApr 26, 2019

The Algorand Foundation will soon launch the Algo from Singapore. At that time, thanks to its innovative technology, the Algorand blockchain will enjoy three important properties.

  1. NO FORKS. Due to its unique consensus algorithm, the Algorand blockchain does not fork. Once it appears, each new block is guaranteed to remain on the chain forever. This property implies that every transaction in the Algorand blockchain is final.
  2. A TRUE MEDIUM OF EXCHANGE. The Algo will be a true medium of exchange. Bitcoin and other proof-of-work projects produce blocks slowly, regardless of how fast the underlying communication network can be. Even if a block could be circulated throughout the network in a few seconds, Bitcoin would still need to generate a block every few minutes in order to keep the chance of a fork sufficiently low. By contrast, because its blockchain does not fork, Algorand produces a block as fast as it can be circulated throughout the network.
  3. UNIVERSAL TOKENIZATION. The Algorand blockchain will enjoy — from day 1— the ability to conduct Dutch auctions, on chain and on a global scale. This will facilitate the tokenization of all kinds of assets — e.g., real estate, distributed computation, etc. — with maximum scalability, fairness, and transparency.

This is just a start. Before I share a few highlights of Algorand’s forthcoming technology, I find it important to explain why these technologies have been prioritized as the most important early advancements to the Algorand blockchain protocol.

Our Development Philosophy

Our development philosophy at Algorand adheres to four fundamental principles. These help us prioritize the innovations we build for a blockchain that aims to provide sustainable and boundless opportunity for all.

  1. PERFORMANCE. Our primary focus early on is ensuring the utmost performance of the Algorand blockchain. Of course, this is a big advantage to algorand users, but is also vital to retain the decentralization of the Algorand blockchain. Centralization has a devious way of creeping in, and one of the main avenues for this type of creep is performance. Put another way, whenever operational requirements become too heavy it is only natural they fall into fewer and fewer hands.
  2. DIRECT USAGE AND ADOPTION. It is not enough to simply develop new tools. There must be a rigorous focus on the specific purpose and utility of the tool being developed as driven by the precise needs of the community. The development of general purpose capabilities is like boiling the ocean in an effort to catch a specific fish. Our efforts focus on building those fundamental layer-1 and layer-2 capabilities that will drive accelerated, broad based adoption of our platform.
  3. INSTANT VERIFIABILITY. Blockchain promised transparency by making information available across the entire chain. However, when data on the chain grows exponentially, transparency erodes leaving critical data hiding in plain sight, lost among the volume of information available across the chain. In particular, sidechains are often implemented by dumping corresponding transactions in general blocks. This obliges the users to scavenge the entire blockchain to reconstruct information relative to the side chain in question. Much like finding a needle in a haystack. Such careful combing of the blockchain may take months, and the time required to verify an assertion greatly diminishes the potential value the blockchain overall.
  4. SIMPLICITY. Underlying all of our technical innovation is the fundamental principle of simplicity. Simplicity is the ultimate sophistication, and is the gateway to broad participation, reliability and trust. To be effective a solution needs to be as simple as possible, but not simpler.

VAULT: ALGORAND’S SOLUTION TO THE BLOCKCHAIN STORAGE PROBLEM

Storing the entire blockchain is necessary in order to retrieve past transactions, as is required in the normal course of business. But for some projects, like Bitcoin, it is also necessary to establish the validity of new transactions.

As a blockchain grows, however, storing all its blocks may prove challenging. Even blockchains that produce a block only every 10 minutes will ultimately consist of more than a terabyte of data. Downloading so much data, in order to catch up to the blockchain, deters new users from actively participating in the generation of new blocks. And relying on other users who store the entire chain to provide you with the occasional past blocks you wish to retrieve is not as simple as it seems. In fact, to avoid being cheated, you would also need to ask for a proof that the information they provide is correct. With currently technology, such a proof may itself be cumbersome, once the blockchain is long enough.

The way traditional blockchains handle the storage and retrieval of blocks does not scale for a decentralized architecture and will ultimately prevent the growth of a borderless economy.

Algorand’s Vault System provides a unique solution to these problems.

  1. Efficient On-Boarding of New Users. Vault makes it vastly simpler for new nodes joining the network to participate in the generation of new blocks. Indeed, Vault allows a new user to download (and to update) just one compact piece of information. So, a new user can participate in the generation of new blocks in exactly the same manner as a user who has been continually participating (and storing all blocks) from day 1.
  2. Efficient and Provable Block Retrieval. Vault enables a user storing the entire blockchain to provably provide the content of any past block (in full or in part). A user who stores the chain can not only provide information relating to past transactions, but also add a short and easily verifiable proof that the information provided is correct.

In sum, Vault safeguards both the decentralization of the Algorand blockchain and the accessibility of the chain to parties who never stored it.

We all know how valuable random-access memory is for computation. Vault will bring the same value to the Algorand blockchain.

PIXEL: Algorand’s Bandwidth Reduction via Better Signatures

To operate in a consensual, decentralized, and secure manner, the Algorand protocol often asks multiple users to sign a common message. (For instance, a new block is added to the blockchain only after it has been signed by multiple and randomly selected verifiers, who have ascertained its validity.)

To reduce the associated bandwidth while keeping the required security, Algorand uses a new digital signature scheme, Pixel.

Pixel-signatures simultaneously enjoy the following two properties:

  1. AGGREGATION: They enable anyone to aggregate multiple signatures of the same message into a single signature of the same length.¹
  2. FORWARD SECURITY. They prevent an adversary, who corrupts a user at a given point in time, from forging the user’s signatures on previous messages.

¹A bit more precisely, let SIGa(M), SIGb(M), … be the digital signatures of M relative to public keys a, b, … Then, Pixel lets one substitute (a,b,…; SIGa(M), SIGb(M), …) with (a,b,… ; SIGa,b,…(M)).

To be sure, aggregatable signatures, such as BLS signatures, already exist. However, their usability is limited due to the lack of a common reference standard. To bridge the gap between theory and implementation, Algorand has put together a standard of BLS signatures that is currently under review by the International Engineering Task Force (IETF). Algorand will use BLS signatures to authenticate individual transactions.

Also, forward-secure signatures already exist in the literature. However, we have been lacking signatures schemes that enjoy both aggregation and forward-security.

Algorand will use Pixel signatures for the entire block generation process. Modern blockchains will be the first — but not the last!— beneficiaries of this innovation.

SELF-VALIDATING TRANSACTIONS: Algorand’s Separation of Consensus and Storage

Algorand’s self-validating transactions is a new technology that finally achieves a clean separation between consensus and storage. Such a separation has been long sought. Achieving it enhances the efficiency and the logical structure of the Algorand blockchain.

Algorand users may certainly check whether payments are valid by constantly maintaining the balance of each public key in the system; that is, how many algos each key has available to spend. As the number of keys grows, however, accurately and efficiently maintaining all balances may become the most demanding computational task for an Algorand node. (Indeed, the other computational requirements are, in the case of our blockchain, quite trivial.)

Algorand’s self-validating payment technology totally dispenses with the need to maintain balances for the purpose of verifying blocks and/or individual payments.

At a very high level, in order to be included in block N + 1, a payment P is propagated together with a very short and easily verifiable proof of its validity, up to — and including — block N. These proofs can (1) be easily and publicly computed given the balances involved in P and (2) be trivially checked by anyone, even those who have no prior knowledge of any current balance in the Algorand system.

Algorand’s self-validating technology extends, beyond payments, to other transactions as well. In particular, it will be crucial to guarantee the efficiency and interoperability of non-fungible tokens in the Algorand blockchain.

ATOMIC² SWAPS: Algorand’s (Truly Atomic) Atomic Swaps

The typical goal of a cross-chain atomic swap is to enable two users of different blockchains to exchange the cryptocurrencies they hold in a peer-to-peer fashion; that is, without the help of any intermediaries. For instance, a user in the Ethereum blockchain may wish to exchange some of his Ether for Algos owned by a user in the Algorand blockchain.

Such swaps aspire to be ‘atomic’: namely, one party cannot get what he wants without the other party getting what she wants. However, they are very far from being atomic in their implementation. In existing blockchains, transactions can be reverted, history can be changed, forked, etc. Hence, existing atomic swaps must rely on complicated multi-step protocols with long delays and cumbersome time locks. A cross-chain atomic swap may indeed take 3 hours to complete.

In Algorand, all transactions are final and instantly confirmed. Hence Algorand can offer cross-chain atomic swaps that are both faster and truly atomic.

But it can do even better.

Atomic swaps are possible between users of the same blockchain. For instance, they may be used to swap a non-fungible token — e.g., a city apartment posted on the chain— for money or for another non-fungible token and possibly additional money.

Today, same-chain atomic swaps could be implemented with potentially fewer steps, but they are still cumbersome, and they still have multiple steps and long time locks. Accordingly, these swaps are still slow and not truly atomic.

Algorand will dramatically change this state of affairs. Indeed, Algorand’s atomic² swaps will implement same-chain atomic swaps in a single transaction! Thus, thanks to Algorand’s instant finality guarantee, these transactions are ‘atomic’ also from an implementation point of view.

Atomic² swaps will provide additional value to the Algorand blockchain. Indeed, they provide the most efficient implementation of the most basic and crucial form of trade.

POST-AND-SALE TRADES: Algorand’s Non-Interactive Trade Mechanism

Traditional trades typically take time. First, two parties must find each other. Then, they negotiate terms. Finally, they need to execute a trade.

Algorand’s atomic² swaps will vastly simply the final execution, but still presuppose that seller and buyer have already found each other (and agreed on a price).

Sometimes, however, one party may want to save herself the time to find a trading partner. For instance, she may advertise a specific good (e.g., a real-estate property) that she owns on the blockchain and offer it for sale at a price chosen by her.

Algorand allows her to post her property for sale, together with a proof of ownership, her chosen price, and an expiration date for her offer. As soon as a buyer spots the offer and agrees to the price, Algorand allows him to make the trade without involving the seller at all.

Algorand essentially provides sellers the option of a non-interactive atomic² swap. As usual, this special swap also continues to guarantee all of Algorand’s security properties — for example, against “double spending”, “uncertainty of ownership”, etc.

Algorand’s Post-and-Sale trades also apply to the case of multiple distinct goods, where the seller names different prices for different subsets of the goods. As soon as a subset is sold, previously offered subsets that include some of the goods that have just been sold are automatically taken off the market.

Finally, Algorand’s Post-and-Sale trades include non-interactive auctions, where the price is determined by the market, not by the seller.

Post-and-Sale trades realize e-commerce on the blockchain and, due to their speed and ease of use, will provide tremendous value to the Algorand blockchain.

NEXT

Algorand’s technology roadmap is deep. The above innovations are just those we plan to add to our blockchain in the first months following launch. Other technologies, just as innovative, will be added next.

In particular, we will add smart² contracts, quantum-resiliency, and more.

IN SUM

Blockchains have largely remained aspirational, but ultimately, are sophisticated technological products. It is only good and proper to raise the bar of our aspirations. But, without sustaining them with proper technology, aspirations risk remaining a pie in the sky.

Algorand has already deployed much innovative technology and will continue to do so. Indeed, Algorand is wholeheartedly committed to research: theory research, economic research, security research, mathematics research, and systems research.

Powered by its general and flexible consensus protocol, the Algorand blockchain can evolve in a consensual manner. It is therefore uniquely poised to take advantage of new technological breakthroughs. Providing these breakthroughs to Algorand’s blockchain and enabling a truly borderless economy is our cherished mission.

We are and will remain open to collaborations with anyone who shares this mission.

--

--

Silvio Micali
Algorand

Silvio Micali is the Founder of @Algorand. He is one of the co-inventors of zero-knowledge proofs and is a Turing Award-Winning MIT professor.