BlockDAG & Kaspa

I recently wrote a lot of indignation about a piece of crap called ~~BlockDAG~~ ScamDAG, which exploits the name of a new technology called BlockDAG. It would be remiss of me not to write about the real, genuine BlockDAG. In this article, I’ll do just that. Moreover, I’ll talk about the best cryptocurrency project, technologically speaking, in my opinion, that I’ve ever seen. I’ll say right away that this article won’t discuss profitability. Of course, I’m interested in this question; it would be great to be a crypto billionaire 😂🤑, but I’m not a scam artist, so you won’t see any price predictions or other economic analysis here.

What problem does BlockDAG solve?

So, let’s begin. Any distributed network has several competing requirements. It must be secure, decentralized, and fast enough to be convenient for everyday use. In traditional blockchains, increased speed often leads to an increase in the number of conflicting blocks, increased node requirements, and decreased network stability.

BlockDAG attempts to solve this problem not by abandoning the secure but slow Proof of Work (PoW) algorithm (for example, in favor of Proof of Stake (PoS)), but by changing the data structure. Instead of forcing miners to build a single linear chain, a special protocol was developed that allows the creation of multiple blocks in parallel, and then consistently organizes them into a special BlockDAG data structure.

Blockchain

Before delving into the technical aspects of BlockDAG, let’s first return to the blockchain. In traditional blockchains, the data structure is a linear chain of blocks (a linked hash list). Each new block is added on top of the previous one, creating a sequential chain. Parallel blocks in such a network can appear due to propagation delays. Consensus ultimately selects a single canonical branch, and competing blocks are excluded from it. The rules are determined by the principles of consensus mechanisms that ensure data consistency among all network participants.

Proof-of-work consensus

The original, well-studied, and secure Proof of Work (PoW) consensus allows network participants (miners) to agree on which transactions are valid and add them to the blockchain (see the longest (labor-intensive) chain rule [1]). This imposes performance limitations: primarily, we are limited by the number of transactions that can be processed per second, which in turn leads to delays and high fees under increased network load. Furthermore, the PoW consensus mechanism requires significant computing resources, as well as time, leading to high energy costs.

How does this work in practice?

The most famous blockchain with a PoW consensus is undoubtedly Bitcoin, which uses Nakomoto consensus [1] a protocol based on proof-of-work and the longest-chain rule to achieve distributed agreement between network nodes. The Bitcoin network works as follows:

Users send transactions, which include data on cryptocurrency transfers from one address to another. These transactions do not immediately become part of the blockchain; they are first collected in a waiting pool called the mempool.
Miners select transactions from the mempool and form a new block. A block contains not only the transactions, but also the hash of the previous block (to ensure the integrity of the entire chain), as well as a special variable called a nonce.
For a block to be accepted by the network, its hash must satisfy a certain condition: it must be less than a predetermined value, which depends on the network difficulty. To find such a hash, miners iterate over the nonce values, changing them, and calculate the block’s hash. This process is called mining. Miners continue calculating until they find a hash that meets the specified difficulty.
Once a miner finds such a hash, they publish their block to the network. Other nodes in the network verify that the hash meets the conditions and that the transactions in the block are valid.
If the block passes verification, it is added to the local copy of the blockchain shared by all network participants. The block is considered confirmed. This process is consensus, confirming that all network participants agree on a single version of the blockchain.
When a new block is added to the blockchain, each subsequent block references the previous block through its hash. This ensures the continuity of the chain and makes it secure from data modification. Changing the data in one block requires recalculating the hashes of all subsequent blocks, which is extremely difficult, if not impossible, as it would require colossal computing power.

Without a doubt, PoW consensus is a secure solution for blockchains. Especially when compared to Proof of Stake (PoS). However, as mentioned earlier, networks currently use a linear chain of blocks. Only one new block can be added at a time the next block builds solely on the previous one. Combined with PoW, such architectures will be inferior to similar PoS-based architectures.

BlockDAG

BlockDAG (Directed Acyclic Graph) is a data structure, a new way to organize a network of transaction blocks. In BlockDAG, new blocks can be added not only on top of the latest block, but also reference several previous blocks, allowing transactions to be processed in parallel. This removes the throughput limitations typical of classic PoW chains. The following figure shows an illustration of BlockDAG from the original paper by Yonatan Sompolinsky, Shai Wyborski, and Aviv Zohar [2].

Example BlockDAG from the original paper [2]

The figure shows that each block references all blocks known to its miner at the time of its creation. Let’s consider what’s happening in the figure using block $H$ as an example: $$\text{past}(H) = \lbrace Genesis, C, D, E \rbrace$$ are blocks that $H$ references directly or indirectly (in this case, indirectly only the $Genesis$ block), and which were provably created before $H$; $$\text{future}(H) = \lbrace J, K, M \rbrace$$ are blocks that reference $H$ directly or indirectly, and which were provably created after $H$; $$\text{anticone}(H) = \lbrace B, F , I , L \rbrace$$ is the ordering between these blocks and $H$ is ambiguous. Reaching consensus on the ordering between blocks and other blocks in their anticone is the main problem we face. $$\text{tips}(G) = \lbrace J , L, M \rbrace$$ are leaf blocks, specifically, blocks with incoming degree $0$; they will be referenced in the header of the next block.

Instead of extending a single chain, as in the first case, the miner references in its new block all blocks of the DAG that the miner observes locally at the time of creating the new block. Furthermore, the miner must broadcast its new block as quickly as possible. These two rules together constitute the DAG mining protocol in PHANTOM (published in [2]).

How GHOSTDAG achieves consensus

PHANTOM describes the general idea of consensus in BlockDAG, and GHOSTDAG is a practical algorithm used in Kaspa. Its goal is to obtain a uniform block ordering from a graph where multiple blocks can be created nearly simultaneously.

Each block has multiple parents. One of these is chosen by the algorithm as the selected parent. It is used to construct the block ordering. The remaining parents and the blocks between them form an additional part of the graph, which the algorithm must process correctly.

An important parameter is $k$, which specifies how many parallel blocks the protocol allows in a given local environment. If two blocks are in each other’s anticone, meaning the ordering between them is initially unknown, GHOSTDAG attempts to determine whether they can simultaneously be part of a consensus set of blocks. This set is called the blue set. Blocks that are not in it are referred to as the red set. The red color does not necessarily mean that the block was created by an attacker or is completely invalid. It is a classification used by the algorithm when constructing the consensus order. After the blue set is selected, blocks are sorted according to deterministic rules. The merge set plays an important role here a set of blocks that were created in parallel to the chosen parent and must be added to the overall order.

As a result, every node in the network gradually gains a consistent view of the order of blocks and transactions. Nodes may see blocks at different times, however, after they propagate through the network, honest participants must agree on the same result. This is what distinguishes the BlockDAG from a simple set of independent parallel chains.

If two transactions attempt to spend the same Unspent Transaction Output (UTXO), the ordering algorithm determines which one appeared first in the consensus history. The second transaction becomes conflicting and cannot be accepted as a normal, valid spend.

How a transaction goes through the Kaspa network

Let’s look at the path of a typical transaction.

The wallet creates a transaction and signs it with its private key. Kaspa uses the UTXO model: coins are represented by unspent outputs from previous transactions.
The signed transaction is sent to network nodes. Nodes verify the signature, the presence of the required UTXOs, and the absence of an attempt to spend the same coins twice.
The miner includes the transaction in a new block. Unlike Bitcoin, multiple miners can create blocks almost simultaneously, and this does not mean that most of the work done will be immediately discarded.
The new block is propagated through the network and becomes part of the BlockDAG. The GHOSTDAG determines its position relative to other blocks.
As new blocks are mined, the transaction’s position becomes increasingly stable. Additional blocks increase the amount of work an attacker would have to perform to change history.

Therefore, Kaspa’s high speed is due to the fact that first and foremost, the way blocks are organized is changing: they can be created in parallel, and then ordered by consensus rules.

What BlockDAG doesn’t solve automatically

BlockDAG isn’t a magical way to achieve infinite throughput. It removes one of the limitations of a linear chain the need to discard most blocks created in parallel. However, other network limitations remain.

Each block still needs to be transmitted to other nodes, its transactions verified, and the necessary network state information stored. The more frequently blocks are created, the greater the requirements for communication bandwidth, processor, RAM, and disk space.

Furthermore, parallel block creation doesn’t mean that any node can infinitely perform all operations simultaneously. If the block generation rate grows faster than the network’s capacity, nodes will begin to see different parts of the DAG, and synchronization time will increase.

The $k$ parameter is also a compromise. A large value allows for more parallel blocks to be considered, but complicates graph processing. A small value simplifies consensus, but limits the number of blocks that can effectively coexist.

Therefore, it’s more accurate to talk about shifting the bottleneck rather than removing restrictions. In a classic blockchain, the main limitation is the sequential chain. In BlockDAG, data distribution, computational load, and node requirements play a major role.

Also, can you show me projects that have solved these problems? 😁

Kaspa

The GHOSTDAG protocol was implemented as the underlying technology of the Kaspa cryptocurrency. Reddit user commun16 has provided a close-up, but very illustrative visualization of the difference between Bitcoin and Kaspa.

Bitcoin and Kaspa difference

Mining and the Role of KAS

Kaspa is not a token running on top of another blockchain, but an independent cryptocurrency network. Its native coin is called KAS. It is used for transfers and fees, and is also paid to miners as a reward for their participation in securing the network.

Kaspa miners use the kHeavyHash algorithm. Like Bitcoin, a miner’s task is to find a value that satisfies the current difficulty requirements of the block hash. The resulting block is then propagated throughout the network and verified by other nodes.

The miner’s reward is important not only for the miner themselves. It creates an economic incentive to maintain the network and continue performing Proof of Work. Therefore, the coin issuance, at this point in the project’s lifespan, is part of the security mechanism, and not simply a way to issue new tokens.

Kaspa’s maximum supply is approximately 28.7 billion KAS. The emission decreases according to a pre-established schedule. (https://www.scribd.com/document/624616354/KASPA-EMISSION-SCHEDULE-1) This allows for the issuance rules to be known in advance and not dependent on the decisions of a single company or fund.

It’s important to distinguish protocol economics from price predictions. The emission schedule, fees, and miner rewards can be studied technically. However, it’s impossible to automatically derive the future market price of KAS from these parameters.

Safety and compromises

Kaspa’s security isn’t based solely on the use of BlockDAG. It depends on Proof of Work, hashrate distribution, node behavior, and the network’s ability to quickly propagate new blocks. Like any PoW network, Kaspa assumes that honest miners control the majority of the computing power. If an attacker gains a significant share of the hashrate, they can temporarily censor transactions, create alternative history, or attempt to double-spend their own funds. However, the economic viability of such attacks is questionable.

Ecosystem

The Kaspa ecosystem consists of several layers. The base layer includes the protocol itself, a full node, mining, and wallets. The next layer consists of explorers, APIs, and indexers that make BlockDAG data easy to view and use by applications. Developer tools and community projects are also worth highlighting. They can use the network for transfers, token issuance, service creation, and data experiments, but are not necessarily part of the official Kaspa team.

The cryptocurrency ecosystem is not as large as other hyped projects, but it is growing rapidly thanks to the community. There’s an official website kaspa.org, where you can find information about mining and mining pools, news, and learn about future plans. Below I’ll list the main categories of tools that you can use right now.

Wallets

There are several wallets to choose from.

Personally, I prefer Command Line Wallet and KDX because they give the user more control. However, the interface itself doesn’t guarantee security: much more important are the software’s source, the method for storing the seed phrase, and the state of the device you’re using.

There’s also the Kaspium wallet for mobile phones.

I can’t say anything about it, since I haven’t used it.

How to get started safely

Choosing between the command line, a desktop application, or a mobile wallet doesn’t in itself make storing coins safe. As I noted earlier, the source from which the wallet was downloaded, how the seed phrase is stored, and the state of the device on which it is running are more important. After creating a new wallet, you should save the seed phrase offline and never send it to other people or websites. For your first transfer, it’s best to use a small amount. First, you should check that the address is correct and that the transaction has actually arrived. After that, you can find the transaction in the explorer by its ID. The explorer is convenient for checking the network status, but it shouldn’t receive the private key or seed phrase.

Those who want to delve deeper into Kaspa’s operation should try running their own node. This allows you to receive data directly from the network and not rely entirely on a third party.

Explorers

There are several blockchain explorers for tracking transactions. These are the official blockchain explorers for cryptocurrency.

The first one Kaspa Explorer.
More powerful kas.fyi. You can explore the KRC-20 token ecosystem here. Available at link.

kas.fyi web interface

Others

There is also an official Wiki. The Wiki is constantly updated with current information and is translated into various languages.

Conclusion

The main idea behind Kaspa is that it attempts to preserve the proven PoW and open participation model, while allowing miners to create blocks in parallel. GHOSTDAG transforms this parallel graph into a consistent order of blocks and transactions. This allows the network to utilize computational work that would often be lost in a traditional blockchain due to competing blocks.

At the same time, Kaspa doesn’t address the fundamental limitations of distributed systems. It still requires an honest majority of the hashrate, rapid data dissemination, available nodes, and correct software implementation. Therefore, my thesis that Kaspa is the best cryptocurrency project relates primarily to a specific combination of properties: Proof of Work, BlockDAG, open launch, and the desire for high speed without switching to Proof of Stake. Whether this will be sufficient for mass adoption is a separate question. But from a technical perspective, Kaspa represents one of the most interesting experiments in scaling Nakamoto Consensus.

Thoughts on a future project

So far, I’ve mostly talked about the technical side of Kaspa. However, technology and market success are not the same thing.

In many ways, Kaspa is developing according to a model reminiscent of early Bitcoin. The project didn’t have an ICO, a large centralized marketing campaign, or a single owner responsible for product promotion. Development largely rests on the shoulders of the community: developers, miners, coin holders, and enthusiasts. For Bitcoin, this approach has proven extremely successful. At the time of its emergence, the cryptocurrency market was completely empty. There were no thousands of competing networks, large funds, professional marketing departments, or projects with huge advertising budgets. Bitcoin was able to achieve a key position simply because it was first and offered a completely new idea. Today, the situation is very different. The cryptocurrency market is overcrowded. It already boasts projects with large teams, significant funding, aggressive marketing, numerous partnerships, and established ecosystems. Users are constantly being offered new tokens, networks, and applications.

And this raises a key question regarding Kaspa’s future. The project’s technology is indeed very interesting, however, the average market participant won’t study the GHOSTDAG design, the properties of the anticone, or the specifics of the Proof of Work implementation in detail. For most people, other things are more important: is the wallet easy to use, can you quickly buy and send coins, are there the necessary applications, how active is the community, and how frequently the project appears in the media (on Elon Musk’s Twitter, LOL).

It turns out that technical appeal alone may not be enough. It creates a foundation, but it alone doesn’t guarantee users. The project also needs to explain this foundation in simple terms, translate it into user-friendly products, and ensure that people use Kaspa not only out of curiosity about the technology but also to solve specific problems. This article is one attempt to do just that. Although, rereading it as I write, I already realize that it’s too complex for the average user.

The project’s future looks both interesting and uncertain to me. Kaspa can maintain Bitcoin’s principles while learning to effectively disseminate information about itself. But if the technical features remain understood only by a small circle of enthusiasts, the project will have a hard time translating its advantages into widespread use.

Sources

[1] Nakamoto S. Bitcoin: A Peer-to-Peer Electronic Cash System, 2008.
[2] Sompolinsky Y., Wyborski S., Zohar A. PHANTOM GHOSTDAG: A Scalable Generalization of Nakamoto Consensus, 2021. 14 p.

Donation

If you found this material useful, I would appreciate your donation in the form of a transfer to Kaspa’s address: kaspa:qpnke7xh9pe3rdf839mm96emy2ymlwdjgtjhvxkjmasshghqar08xp3ydwhe0.