CS120: Bitcoin for Developers I

Because the blockchain is a decentralized data structure, different copies of it are not always consistent. Blocks might arrive at different nodes at different times, causing the nodes to have different perspectives of the blockchain. To resolve this, each node always selects and attempts to extend the chain of blocks that represents the most Proof-of-Work, also known as the longest chain or greatest cumulative work chain. By summing the work recorded in each block in a chain, a node can calculate the total amount of work that has been expended to create that chain. As long as all nodes select the greatest-cumulative-work chain, the global bitcoin network eventually converges to a consistent state. Forks occur as temporary inconsistencies between versions of the blockchain, which are resolved by eventual reconvergence as more blocks are added to one of the forks.

Tip: The blockchain forks described in this section occur naturally (accidentally) as a result of transmission delays in the global network. Later in this chapter, we will also look at deliberately induced forks (hard forks and soft forks), which are used to modify the consensus rules.

Each node has its own perspective of the global blockchain. As each node receives blocks from its neighbors, it updates its own copy of the blockchain, selecting the greatest-cumulative-work chain. For illustration purposes, each node contains a shape that represents the block that it believes is currently the tip of the main chain. So, if you see a star shape in the node, that means that the star block is the tip of the main chain, as far as that node is concerned.

In the first diagram (Before the fork – all nodes have the same perspective), the network has a unified perspective of the blockchain, with the star block as the tip of the main chain.

Figure 2. Before the fork – all nodes have the same perspective

A "fork" occurs whenever there are two different valid blocks at the same block height competing to form the longest blockchain. This occurs under normal conditions whenever two miners solve the Proof-of-Work algorithm within a short period of time from each other. As both miners discover a solution for their respective candidate blocks, they immediately broadcast their own "winning" block to their immediate neighbors who begin propagating the block across the network. Each node that receives a valid block will incorporate it into its blockchain, extending the blockchain by one block. If that node later sees another valid block extending the same parent (at the same block height), it connects the second block on a secondary chain, forking its main chain. As a result, some nodes will "see" one winning block first, while other nodes will see the other winning block first, and two competing versions of the blockchain will emerge.

In Visualization of a blockchain fork event: two blocks found simultaneously, we see two miners (Node X and Node Y) who mine two different blocks almost simultaneously. Both of these blocks are children of the star block, and extend the chain by building on top of the star block. To help us track it, one is visualized as a triangle block originating from Node X, and the other is shown as an upside-down triangle block originating from Node Y.

Figure 3. Visualization of a blockchain fork event: two blocks found simultaneously

Let's assume, for example, that the miner Node X finds a Proof-of-Work solution for a block "triangle" that extends the blockchain, building on top of the parent block "star". Almost simultaneously, the miner Node Y who was also extending the chain from block "star" finds a solution for block "upside-down triangle," his candidate block. Now, there are two possible blocks; one we call "triangle," originating in Node X; and one we call "upside-down triangle," originating in Node Y. Both blocks were successfully mined, both blocks are valid (contain a valid solution to the Proof-of-Work), and both blocks extend the same parent (block "star"). Both blocks likely contain most of the same transactions, with only perhaps a few differences in the order of transactions.

As the two blocks propagate, some nodes receive block "triangle" first and some receive block "upside-down triangle" first. As shown in Visualization of a blockchain fork event: two blocks propagate, splitting the network, the network splits into two different perspectives of the blockchain; one side topped with the triangle block, the other with the upside-down-triangle block.

Figure 4. Visualization of a blockchain fork event: two blocks propagate, splitting the network

In Visualization of a blockchain fork event: two blocks propagate, splitting the network, the miner Node X mined (created) the triangle block and extended the star chain with it. Therefore, Node X initially considers the chain with "triangle" block as the main chain. Later, Node X also received the "upside-down triangle" block that was mined by Node Y. Since it was received second, it is assumed to have "lost" the race. Yet, the "upside-down triangle" block is not discarded. It is linked to the "star" block parent and forms a secondary chain. While Node X assumes its main chain is the winning chain, it keeps the "losing" chain so that it has the information needed to reconverge if the "losing" chain ends up "winning".

On the other side of the network, the miner Node Y constructs a blockchain based on its own perspective of the sequence of events. The miner Node Y mined (created) the "upside-down triangle" and initially considers that chain as the main chain (the "winner" chain). When it later received the "triangle" block that was mined by Node X, it connected it to the "star" block parent as a secondary chain.

Neither side is "correct," or "incorrect". Both are valid perspectives of the blockchain. Only in hindsight will one prevail, based on how these two competing chains are extended by additional work.

Each mining node whose perspective resembles Node X will immediately begin mining a candidate block that extends the chain with "triangle" as its tip. By linking "triangle" as the parent of their candidate blocks, they are voting with their hashing power. Their vote supports the chain that they have elected as the main chain.

Any mining node whose perspective resembles Node Y will start building a candidate block with "upside-down triangle" as its parent, extending the chain that they believe is the main chain. And so, the race begins again.

Forks are almost always resolved within one block time (10 minutes on average). While part of the network's hashing power is dedicated to building on top of "triangle" as the parent, another part of the hashing power is focused on building on top of "upside-down triangle". Even if the hashing power is almost evenly split, it is likely that one set of miners will find a solution and propagate it before the other set of miners have found any solutions. Let's say, for example, that the miners building on top of "triangle" find a new block "rhombus" that extends the chain (e.g., star-triangle-rhombus). They immediately propagate this new block and the entire network sees it as a valid solution as shown in Visualization of a blockchain fork event: a new block extends one fork, reconverging the network. Both Node X and Node Y now consider "upside-down-triangle" block as a stale block.

All nodes that had chosen "triangle" as the winner in the previous round will simply extend the chain one more block. The nodes that chose "upside-down triangle" as the winner, however, will now see two chains: star-triangle-rhombus and star-upside-down-triangle. The chain star-triangle-rhombus is now longer (more cumulative work) than the other chain. As a result, those nodes will set the chain star-triangle-rhombus as the main chain and change the star-upside-down-triangle chain to a secondary chain, as shown in Visualization of a blockchain fork event: the network reconverges on a new longest chain. This is a chain reconvergence, because those nodes are forced to revise their view of the blockchain to incorporate the new evidence of a longer chain. Any miners working on extending the chain star-upside-down-triangle will now stop that work because their candidate block is now considered a child of a stale block, as its parent "upside-down-triangle" is no longer on the longest chain. Since the upside-down-triangle block is now obsolete, the miner Node Y (which mined this block) will not be able to spend the mining reward for this block, even though this block was valid and was successfully mined. The transactions within "upside-down-triangle" that are not within "triangle" are re-inserted in the mempool for inclusion in the next block to become a part of the main chain. The entire network reconverges on a single blockchain star-triangle-rhombus, with "rhombus" as the last block in the chain. All miners immediately start working on candidate blocks that reference "rhombus" as their parent to extend the star-triangle-rhombus chain.

Figure 5. Visualization of a blockchain fork event: a new block extends one fork, reconverging the network. Both Node X and Node Y now consider "upside-down-triangle" block as a stale block.

Figure 6. Visualization of a blockchain fork event: the network reconverges on a new longest chain

It is theoretically possible for a fork to extend to two blocks, if two blocks are found almost simultaneously by miners on opposite "sides" of a previous fork. However, the chance of that happening is very low. Whereas a one-block fork might occur every day, a two-block fork occurs at most once every few weeks.

Bitcoin's block interval of 10 minutes is a design compromise between fast confirmation times (settlement of transactions) and the probability of a fork. A faster block time would make transactions clear faster but lead to more frequent blockchain forks, whereas a slower block time would decrease the number of forks but make settlement slower.