Bitcoin is structured as a peer-to-peer network architecture on top of the Internet. The term peer-to-peer or P2P means that the computers that participate in the network are peers to each other, they are all equal there are no "special" nodes and all nodes share the burden of providing network services. The network nodes interconnect in a mesh network with a "flat" topology. There is no "server", no centralized service and no hierarchy within the network. Nodes in a peer-to-peer network both provide and consume services at the same time, with reciprocity acting as the incentive for participation. Peer-to-peer networks are inherently resilient, de-centralized and open. The pre-eminent example of a P2P network architecture was the early Internet itself, where nodes on the IP network were equal. Today's Internet architecture is more hierarchical, but the Internet Protocol still retains its flat-topology essence. Beyond bitcoin, the largest and most successful application of P2P technologies is file sharing, with Napster as the pioneer and bittorrent as the most recent evolution of the architecture.
Bitcoins P2P network architecture is much more than a topology choice. Bitcoin is a peer-to-peer digital cash system by design, and the network architecture is both a reflection and a foundation of that core characteristic. De-centralization of control is a core design principle and that can only be achieved and maintained by a flat, de-centralized P2P consensus network.
The term "bitcoin network" refers to the collection of nodes running the bitcoin P2P protocol. In addition to the bitcoin P2P protocol, there are other protocols such as Stratum, that are used for mining and lightweight or mobile wallets. These additional protocols are provided by gateway routing servers that access the bitcoin network using the bitcoin P2P protocol and then extend that network to nodes running other protocols. For example, Stratum servers connect Stratum mining nodes via the Stratum protocol to the main bitcoin network and bridge the Stratum protocol to the bitcoin P2P protocol. We use the term "extended bitcoin network" to refer to the overall network that includes the bitcoin P2P protocol, pool mining protocols, the Stratum protocol and any other related protocols connecting the components of the bitcoin system.
While nodes in the bitcoin P2P network are equal, they may take on different "roles", depending on the functionality they are supporting. A bitcoin node is a collection of functions: routing, the blockchain database, mining, and wallet services. A full node with all four of these functions is shown below:
[[full_node_reference]]
.A bitcoin network node with all four functions: Network routing, Blockchain database, Mining and Wallet
All nodes include the routing function to participate in the network and may include other functionality. All nodes validate and propagate transactions and blocks, discover and maintain connections to peers. In the full node example above, the routing function is indicated by an orange circle named "Network Routing Node".
Some nodes, called full nodes, also maintain a complete and up-to-date copy of the blockchain. Full nodes can autonomously and authoritatively verify any transaction without external reference. Some nodes maintain only a subset of the blockchain and verify transactions using a method called _Simple Payment Verification_ or SPV. These nodes are known as SPV or Lightweight nodes. In the full node example above, the full node blockchain database function is indicated by a blue circle named "Blockchain Database". SPV nodes are drawn without the blue circle, showing that they do not have a full copy of the blockchain.
Mining nodes compete to create new blocks by running specialized hardware to solve the proof-of-work algorithm. Some mining nodes are also full nodes, maintaining a full copy of the blockchain while others are lightweight nodes participating in pool mining and depending on a pool server to maintain a full node. The mining function is shown in the full node above as a black circle named "Mining".
User wallets may be part of a full node, as is usually the case with desktop bitcoin clients. Increasingly many user wallets, especially those running on resource constrained devices such as smart phones, are SPV nodes. The wallet function is shown above as a green circle named "Wallet".
In addition to the main node types on the bitcoin P2P protocol, there are servers and nodes running other protocols, such as specialized mining pool protocols and lightweight client access protocols.
Here are the most common node types on the extended bitcoin network:
[[node_type_ledgend]]
.Different types of nodes on the extended bitcoin network
The main bitcoin network, running the bitcoin P2P protocol consists of between 7,000 to 10,000 nodes running various versions of the bitcoin reference client (Bitcoin Core) and a few hundred nodes running various other implementations of the bitcoin P2P protocol, such as BitcoinJ, Libbitcoin and btcd. A small percentage of the nodes on the bitcoin P2P network are also mining nodes, competing in the mining process, validating transactions and creating new blocks. Various large companies interface with the bitcoin network by running full-node clients based on the Bitcoin Core client, with full copies of the blockchain and a network node, but without mining or wallet functions. These nodes act as network edge routers, allowing various other services (exchanges, wallets, block explorers, merchant payment processing) to be built on top.
The extended bitcoin network includes the network running the bitcoin P2P protocol, described above, as well as nodes running specialized protocols. Attached to the main bitcoin P2P network are a number of pool servers and protocol gateways that connect nodes running other protocols, mostly pool mining nodes (see <<mining>>) and lightweight wallet clients, which do not carry a full copy of the blockchain.
The diagram below shows the extended bitcoin network with the various types of nodes, gateway servers, edge routers and wallet clients and the various protocols they use to connect to each other.
[[bitcoin_network]]
.The extended bitcoin network showing various node types, gateways and protocols
When a new node boots up, it must discover other bitcoin nodes on the network in order to participate. To start this process, a new node must discover at least one existing node on the network and connect to it. The geographic location of the other nodes is irrelevant, the bitcoin network topology is not geographically defined. Therefore, any existing bitcoin nodes can be selected at random.
To connect to a known peer, nodes establish a TCP connection, usually to port 8333 (the bitcoin "well known" port), or an alternative port if one is provided. Upon establishing a connection, the node will start a "handshake" by transmitting a +version+ message, which contains basic identifying information, including:
(See https://github.com/bitcoin/bitcoin/blob/d3cb2b8acfce36d359262b4afd7e7235eff106b0/src/net.cpp#L562 for an example of the +version+ network message)
The peer node responds with +verack+ to acknowledge and establish a connection, and optionally sends its own +version+ message if it wishes to reciprocate the connection and connect back as a peer.
How does a new node find peers? While there are no special nodes in bitcoin, there are some long running stable nodes that are listed in the client as _seed nodes_. While a new node does not have to connect with the seed nodes, it can use them to quickly discover other nodes in the network. In the Bitcoin Core client, the option to use the seed nodes is controlled by the option switch +-dnsseed+, which is set to 1, to use the seed nodes, by default. Alternatively, a bootstrapping node that knows nothing of the network must be given the IP address of at least one bitcoin node after which it can establish connections through further introductions. The command line argument +-seednode+ can be used to connect to one node just for introductions, using it as a DNS seed. After the initial seed node is used to form introductions, the client will disconnect from it and use the newly discovered peers.
Once one or more connections is established, the new node will send an +addr+ message containing its own IP address, to its neighbors. The neighbors will in turn forward the +addr+ message to their neighbors, ensuring that the newly connected node becomes well known and better connected. Additionally, the newly connected node can send +getaddr+ to the neighbors asking them to return a list of IP addresses of other peers. That way, a node can find peers to connect to and advertise its existence on the network for other nodes to find it. On a node running the Bitcoin Core client, you can list the peer connections with the command +getpeerinfo+:
A node must connect to a few different peers in order to establish diverse paths into the bitcoin network. These paths are not reliable, nodes come and go, and so the node must continue to discover new nodes as it loses old connections as well as assist other nodes when they bootstrap. Only one connection is needed to bootstrap, as the first node can offer introductions to its peer nodes and those peers can offer further introductions. Its also unnecessary and wasteful of network resources to connect to more than a handful of nodes. After bootstrapping a node will remember its most recent successful peer connections, so that if it is rebooted it can quickly reestablish connections with its former peer network. If none of the former peers respond to its connection request, the node can use the seed nodes to bootstrap again.
To override the automatic management of peers and to specify a list of IP addresses, users can provide the option +-connect=<IPAddress>+ and specify one or more IP addresses. If this option is used, the node will only connect to the selected IP addresses, instead of discovering and maintaining the peer connections automatically.
If there is no traffic on a connection, nodes will periodically send a message to maintain the connection. If a node has not communicated on a connection for more than 90 minutes it is assumed to be disconnected and a new peer will be sought. Thus the network dynamically adjusts to transient nodes, network problems, and can organically grow and shrink as needed without any central control.
Full nodes are nodes that maintain a full blockchain, with all transactions. More accurately they probably should be called "full blockchain nodes". In the early years of bitcoin, all nodes were full nodes and currently the Bitcoin Core client is a full blockchain node. In the last two years however, new forms of bitcoin clients have been introduced, which do not maintain a full blockchain but run as lightweight clients. These are examined in more detail in the next section.
Full blockchain nodes maintain a complete and up-to-date copy of the bitcoin blockchain, with all the transactions, which they independently build and verify, starting with the very first block (genesis block) and up to the latest known block in the network. A full blockchain node can independently and authoritatively verify any transaction, without recourse or reliance on any other node or source of information. The full blockchain node relies on the network to receive updates about new blocks of transactions, which it then verifies and incorporates into its local copy of the blockchain.
Running a full blockchain node gives you the pure bitcoin experience: independent verification of all transactions without the need to rely on, or trust, any other systems. It's easy to tell if you're running a full node because it requires several gigabytes of persistent storage (disk space) to store the full blockchain. If you need a lot of disk and it takes 2-3 days to "sync" to the network you are running a full node. That is the price of complete independence and freedom from central authority.
There are a few alternative implementations of full-blockchain bitcoin clients, built using different programming languages and software architectures. However, the most common implementation is the reference client Bitcoin Core, also known as the Satoshi Client. More than 90% of the nodes on the bitcoin network run various versions of Bitcoin Core. It is identified as "Satoshi" in the sub-version string sent in the +version+ message and shown by the command +getpeerinfo+ as we saw above, for example +/Satoshi:0.8.6/+.
=== Exchanging "Inventory"
The first thing a full node will do once it connects to peers is try to construct a complete blockchain. If it is a brand-new node and has no blockchain at all, then it only knows one block (the genesis block), which is statically embedded in the client software. Starting with block #0, the genesis block, the new node will have to download hundreds of thousands of blocks to synchronize with the network and establish a full blockchain.
The process of "syncing" the blockchain starts with the +version+ message, as that contains +BestHeight+, a node's current blockchain height (number of blocks). A node will see the +version+ messages from its peers, know how many blocks they each have and be able to compare to how many blocks it has in its own blockchain. Peered nodes will exchange a +getblocks+ message that contains the hash (fingerprint) of the top block on their local blockchain. One of the peers will be able to identify the received hash as belonging to a block that is not at the top, but rather belongs to an older block, thus deducing that its own local blockchain is longer than its peer's.
The peer that has the longer blockchain has more blocks that the other node and can identify which blocks the other node needs to "catch up". It will identify the first 500 blocks to share and transmit their hashes using an +inv+ (inventory) message. The node missing these blocks will then retrieve them, by issuing a series of +getdata+ messages requesting the full block data and identifying the requested blocks using the hashes from the +inv+ message.
Let's assume for example that a node only has the genesis block. It will then receive an +inv+ message from its peers containing the hashes of the next 500 blocks in the chain. It will start requesting blocks from all its connected peers, spreading the load and ensuring that it doesn't overwhelm any peer with requests. The node keeps track of how many blocks are "in transit" per peer connection, meaning blocks that it has requested but not received, checking that it does not exceed a limit (MAX_BLOCKS_IN_TRANSIT_PER_PEER). This way, if it needs a lot of blocks, it will only request new ones as previous requests are fulfilled, allowing the peers to control the pace of updates and not overwhelming the network. As each block is received, it is added to the blockchain as we will see in the next chapter <<blockchain>>. The local blockchain is gradually built up, more blocks are requested and received and the process continues until the node catches up to the rest of the network.
This process of comparing the local blockchain with the peers and retrieving any missing blocks happens any time a node goes offline for any period of time. Whether a node has been offline for a few minutes and is missing a few blocks, or a month and is missing a few thousand blocks, it starts by sending +getblocks+, gets an +inv+ response and starts downloading the missing blocks.
Not all nodes have the ability to store the full blockchain. Many bitcoin clients are designed to run on space- and power-constrained devices, such as smartphones, tablets or embedded systems. For such devices, a _simple payment verification_ (SPV) method is used to allow them to operate without storing the full blockchain. These types of clients are called SPV clients or lightweight clients. As bitcoin adoption surges, the SPV node is becoming the most common form of bitcoin node, especially for bitcoin wallets.
SPV nodes download only the block headers and do not download the transactions included in each block. The resulting chain of blocks, without transactions, is 1,000 times smaller than the full blockchain. SPV nodes cannot construct a full picture of all the UTXO that is available for spending, as they do not know about all the transactions on the network. SPV nodes verify transactions using a slightly different methodology that relies on peers to provide partial views of relevant parts of the blockchain on-demand.
As an analogy, a full node is like a tourist in a strange city, equipped with a detailed map of every street and every address. By comparison, an SPV node is like a tourist in a strange city asking random strangers for turn-by-turn directions while knowing only one main avenue. While both tourists can verify the existence of a street by visiting it, the tourist without a map doesn't know what lies down any of the side streets and doesn't know what other streets exist. Positioned in front of 23 Church Street, the tourist without a map cannot know if there are a dozen other "23 Church Street" addresses in the city and whether this is the right one. The map-less tourist's best chance is to ask enough people and hope some of them are not trying to mug the tourist.
Simple Payment Verification verifies transactions by reference to their _depth_ in the blockchain instead of their _height_. Whereas a full-blockchain node will construct a fully verified chain of thousands of blocks and transactions reaching down the blockchain (back in time) all the way to the genesis block, an SPV node will verify the chain of all blocks and link that chain to the transaction of interest.
For example, when examining a transaction in block 300,000, a full node links all 300,000 blocks down to the genesis block and builds a full database of UTXO, establishing the validity of the transaction by confirming that the UTXO remains unspent. An SPV node cannot validate whether the UTXO is unspent. Instead, the SPV node will establish a link between the transaction and the block that contains it, using a Merkle Path (see <<merkle_trees>>). Then, the SPV node waits until is sees the six blocks 300,001 through 300,006 piled on top of the block containing the transaction and verifies it by establishing its depth under blocks 300,006 to 300,001. The fact that other nodes on the network accepted block 300,000 and then did the necessary work to produce 6 more blocks on top of it is proof, by proxy, that the transaction was not a double-spend.
An SPV node cannot be persuaded that a transaction exists in a block, when it does not in fact exist. The SPV node establishes the existence of a transaction in a block by requesting a merkle path proof and by validating the proof-of-work in the chain of blocks. However, a transaction's existence can be "hidden" from an SPV node. An SPV node can definitely prove that a transaction exists but cannot verify that a transaction, such as a double-spend of the same UTXO, doesn't exist because it doesn't have a record of all transactions. This type of attack can be used as a Denial-of-Service attack or as a double-spending attack against SPV nodes. To defend against this, an SPV node needs to connect randomly to several nodes, to increase the probability that it is in contact with at least one honest node. SPV nodes are therefore vulnerable to network partitioning attacks or Sybil attacks, where they are connected to fake nodes or fake networks and do not have access to honest nodes or the real bitcoin network.
For most practical purposes, well-connected SPV nodes are secure enough, striking the right balance between resource needs, practicality and security. For the truly security conscious, however, nothing beats running a full blockchain node.
A full blockchain node verifies a transaction by checking the chain of thousands of blocks below it and checks the UTXO is not spent, whereas an SPV node checks how deep the block is buried by a handful of blocks above it.
To get the block headers, SPV nodes use a +getheaders+ message instead of +getblocks+. The responding peer will send up to 2000 block headers using a single +headers+ message. The process is otherwise the same as that used by a full node to retrieve full blocks. SPV nodes also set a filter on the connection to peers, to filter the stream of future blocks and transactions sent by the peers. Any transactions of interest are retrieved using a +getdata+ request. The peer generates a +tx+ message containing the transactions, in response.
Because SPV nodes need to retrieve specific transactions in order to selectively verify them, they also create a privacy risk. Unlike full-blockchain nodes, which collect all transactions within each block, the SPV node's requests for specific data can inadvertently reveal the addresses in their wallet. For example, a third party monitoring a network could keep track of all the transactions requested by a wallet on an SPV node and use those to associate bitcoin addresses with the user of that wallet, destroying the user's privacy.
Shortly after the introduction of SPV/lightweight nodes, the bitcoin developers added a feature called _bloom filters_ to address the privacy risks of SPV nodes. Bloom filters allow SPV nodes to receive a subset of the transactions without revealing precisely which addresses they are interested in, through a filtering mechanism that uses probabilities rather than fixed patterns.
A bloom filter is a probabilistic search filter, a way to describe a desired pattern without specifying it exactly. Bloom filters offer an efficient way to express a search pattern while protecting privacy. They are used by SPV nodes to ask their peers for transactions matching a specific pattern, without revealing exactly which addresses they are searching for.
Using our previous analogy of a tourist without a map asking for directions to a specific address "23 Church St". If they ask strangers for directions to this street, they inadvertently reveal their destination. A bloom filter is like asking "Are there any streets in this neighborhood whose name ends in R-C-H". A question like that reveals slightly less about the desired destination, than asking for "23 Church St". Using this technique, a tourist could specify the desired address in more detail as "ending in U-R-C-H" or less detail as "ending in H". By varying the precision of the search, the tourist reveals more or less information, at the expense of getting more or less specific results. If they ask a less specific pattern, they get a lot more possible addresses and better privacy but many of the results are irrelevant. If they ask for a very specific pattern then they get fewer results but they lose privacy.
Bloom filters serve this function by allowing an SPV node to specify a search pattern for transactions that can be tuned towards precision or privacy. A more specific bloom filter will produce accurate results, but at the expense of revealing what addresses are used in the user's wallet. A less specific bloom filter will produce more data about more transactions, many irrelevant to the node, but will allow the node to maintain better privacy.
An SPV node will initialize a bloom filter as "empty" and in that state the bloom filter will not match any patterns. The SPV node will then make a list of all the addresses in its wallet and create a search pattern matching the transaction output that corresponds to each address. Usually, the search pattern is a Pay-to-Public-Key-Hash script that is the expected locking script that will be present in any transaction paying to the public-key-hash (address). If the SPV node is tracking the balance of a P2SH address, then the search pattern will be a Pay-to-Script-Hash script, instead. The SPV node then adds each of the search patterns to the bloom filter, so that the bloom filter can recognize the search pattern if it is present in a transaction. Finally, the bloom filter is sent to the peer and the peer uses it to match transactions for transmission to the SPV node.
Bloom filters are implemented as a variable-size array of N binary digits (a bit field) and a variable number of M hash functions. The hash functions are designed to always produce an output that is between 1 and N, corresponding to the array of binary digits. The hash functions are generated deterministically, so that any node implementing a bloom filter will always use the same hash functions and get the same results for a specific input. By choosing different length (N) bloom filters and a different number (M) of hash functions, the bloom filter can be tuned, varying the level of accuracy and therefore privacy. In the example below, we use a very small array of 16 bits and a set of 3 hash functions to demonstrate how bloom filters work.
[[bloom1]]
.An example of a simplistic bloom filter, with 16 bit field and 3 hash functions
image::images/Bloom1.png["Bloom1"]
The bloom filter is initialized so that the array of bits is all zeros. To add a pattern to the bloom filter, the pattern is hashed by each hash function in turn. Applying the first hash function to the input results in a number between 1 and N. The corresponding bit in the array (indexed from 1 to N) is found and set to +1+, thereby recording the output of the hash function. Then, the next hash function is used to set another bit and so on and so forth. Once all M hash functions have been applied, the search pattern will be "recorded" in the bloom filter as M bits have been changed from +0+ to +1+.
Here's an example of adding a pattern "A" to the simple bloom filter shown above:
[[bloom2]]
.Adding a pattern "A" to our simple bloom filter
image::images/Bloom2.png["Bloom2"]
Adding a second pattern is as simple as repeating this process. The pattern is hashed by each hash function in turn and the result is recorded by setting the bits to +1+. Note that as a bloom filter is filled with more patterns, a hash function result may coincide with a bit that is already set to +1+ in which case the bit is not changed. In essence, as more patterns record on overlapping bits, the bloom filter starts to become saturated with more bits set to +1+ and the accuracy of the filter decreases. This is why the filter is a probabilistic data structure - it gets less accurate as more patterns are added. The accuracy depends on the number of patterns added versus the size of the bit array (N) and number of hash functions (M). A larger bit array and more hash functions can record more patterns with higher accuracy. A smaller bit array or fewer hash functions will record fewer patterns and produce less accuracy.
Below is an example of adding a second pattern "B" to the simple bloom filter:
[[bloom3]]
.Adding a second pattern "B" to our simple bloom filter
image::images/Bloom3.png["Bloom3"]
To test if a pattern is part of a bloom filter, the pattern is hashed by each hash function and the resulting bit pattern is tested against the bit array. If all the bits indexed by the hash functions are set to +1+, then the patten is _probably_ recorded in the bloom filter. Since the bits may be set because of overlap from multiple patterns, the answer is not certain, but is rather probabilistic. In simple terms, a bloom filter positive match is a "Maybe, Yes".
Below is an example of testing the existence of pattern "X" in the simple bloom filter. The corresponding bits are set to +1+, so the pattern is probably a match:
[[bloom4]]
.Testing the existence of pattern "X" in the bloom filter. The result is probabilistic positive match, meaning "Maybe"
image::images/Bloom4.png["Bloom4"]
On the contrary, if a pattern is tested against the bloom filter and any one of the bits is set to +0+, then this proves that the pattern was not recorded in the bloom filter. A negative result is not a probability, it is a certainty. In simple terms, a negative match on a bloom filter is a "Definitely No".
Below is an example of testing the existence of pattern "Y" in the simple bloom filter. One of the corresponding bits is set to +0+, so the pattern is definitely not a match:
Bloom filters are used to filter the transactions (and blocks containing them) that an SPV node receives from its peers. SPV nodes will create a filter that matches only the addresses held in the SPV node's wallet. The SPV node will then send a +filterload+ message to the peer, containing the bloom filter to use on the connection. After a filter is established, the peer will then test each transaction's outputs against the bloom filter. Only transactions which match the filter are sent to the node.
In response to a +getdata+ message from the node, peers will send a +merkleblock+ message that contains only block headers for blocks matching the filter and a merkle path (See <<merkle_trees>>) for each matching transaction. The peer will also then send +tx+ messages containing the transactions matched by the filter.
The node setting the bloom filter can interactively add patterns to the filter by sending a +filteradd+ message. To clear the bloom filter, the node can send a +filterclear+ message. Since it is not possible to remove a pattern from a bloom filter, a node has to clear and re-send a new bloom filter if a pattern is no longer desired.
Almost every node on the bitcoin network maintains a temporary list of unconfirmed transactions called the memory pool or transaction pool. Nodes use this pool to keep track of transactions that are known to the network but are not yet included in the blockchain. For example, a node that holds a users wallet will use the transaction pool to track incoming payments to the users wallet that have been received on the network but are not yet confirmed.
Some node implementations also maintain a separate pool of orphaned transactions as detailed in <<orphan_transactions>>. If a transaction's inputs refer to a transaction that is not yet known, a missing parent, then the orphan transaction will be stored temporarily in the orphan pool until the parent transaction arrives.
When a transaction is added to the transaction pool, the orphan pool is checked for any orphans that reference this transaction's outputs (its children). Any orphans found are pulled from the orphan pool and validated using the above checklist. If valid, they are also added to the transaction pool, completing the chain that started with the parent transaction. In light of the newly added transaction which is no longer an orphan, the process is repeated recursively looking for any further descendants, until no more descendants are found. Through this process, the arrival of a parent transaction triggers a cascade reconstruction of an entire chain of interdependent transactions by re-uniting the orphans with their parents all the way down the chain.
Both the transaction pool and orphan pool (where implemented) are stored in local memory and are not saved on persistent storage, rather they are dynamically populated from incoming network messages. When a node starts, both pools are empty and are gradually populated with new transactions received on the network.
Some implementations of the bitcoin client also maintain a UTXO database or UTXO pool which is the set of all unspent outputs on the blockchain. While the name "UTXO pool" sounds similar to the transaction pool, it represents a different set of data. Unlike the transaction and orphan pools, the UTXO pool is not initialized empty but instead contains millions of entries of unspent transaction outputs including some dating back to 2009. The UTXO pool may be housed in local memory or as an indexed database table on persistent storage.
Whereas the transaction and orphan pools represent a single node's local perspective and may vary significantly from node to node depending upon when the node was started or restarted, the UTXO pool represents the emergent consensus of the network and therefore will vary little between nodes. Furthermore the transaction and orphan pools only contain unconfirmed transactions, while the UTXO pool only contains confirmed outputs.
