Source | Hackernoon
Compilation | Fire Sauce
Editor | Carol
Produced | Blockchain Camp (blockchain_camp)
Centralized exchanges are arguably one of the most difficult parts of the crypto asset market to analyze . Although the behaviors of many participants in the encryption field are transparently recorded in distributed ledgers, centralized exchanges are still largely operated off-chain, only publishing a subset of their activities to the corresponding blockchain on.
There is no doubt that centralized exchanges will bring a certain degree of opacity and even pose challenges to the most complex analysis techniques.
However, the analysis of the behavior of centralized exchanges can bring many interesting benefits to crypto investors and traders .
Imagine if you could track large cryptographic transmissions between transactions and predict the location of a large number of specific crypto assets, wouldn't it be beautiful? However, all of this requires an understanding of the basic models in centralized exchanges.
The author of this article has been actively researching truly advanced machine learning models to help us understand the behavior of known participants in the crypto field, including centralized exchanges.
Although the machine learning model is quite mature, in many cases, we still need to rely on human-centered data exploration to verify some results.
To meet this challenge, Pablo Biancotto , technical director of IntoTheBlock, led the implementation of an internal chart browser tool that can visualize the topological map of the relationship between addresses , exchanges, and other known entities in the blockchain .
Let beginners vividly and intuitively understand the elegant models that occur every day in centralized exchanges.
Center of the Study of Transaction
To understand the behavior of centralized crypto exchanges, you need to analyze some key components individually or in groups, and generalize some concepts at a higher level.
Here are a few key components related to the behavior of centralized crypto exchanges:
Hot wallet:
Hot wallet usually refers to the main communication method between external parties and exchanges. This wallet can be used to trade cryptocurrency assets.
Cold wallet: A
cold wallet is seen as a means of secure storage of encrypted assets. Such wallets usually hold large amounts of assets that are not intended for frequent transactions.
Deposit address:
usually refers to a temporary address on the blockchain used to transfer funds to the exchange. The main purpose of this type of address is to facilitate users to exchange cash flow.
Withdrawal address:
usually refers to a temporary address on the blockchain used to transfer funds from the main transaction wallet. Sometimes the withdrawal address can also play the dual role of the deposit address.
Although these four components represent the core of the on-chain architecture of a centralized exchange, it requires quite complex heuristics or machine learning methods to identify them. Part of the challenge is that the interaction model between these components can be arbitrary and complex, and not well understood.
With the help of
IntoTheBlock
's transaction machine learning classifier, the visual exploration of the blockchain data set revealed some very elegant and charming models.
Visualize crypto transactions in the blockchain
In order to better understand the behavior of the key components of a centralized crypto exchange, we first explain some of the visual components that highlight their key interaction modes.
1. The deposit address transfers funds to the main transaction wallet
One of the main models in crypto exchange is to transfer funds from the deposit address to the main transaction wallet. The following figure clearly illustrates the model, the blue circle represents the deposit address, and the green circle represents the main wallet. It should be noted how many deposit addresses are bound to the main wallet in a single transaction.
The figure below shows the same model in a larger scale:
2. The main wallet allocates funds to the withdrawal address
Another traditional model in centralized exchanges is to send transactions to the withdrawal address. The following figure visually demonstrates this model. The orange circle represents the withdrawal address and the blue circle represents the main transaction wallet.
3. The withdrawal address also serves as the deposit address
In some cases, the withdrawal address will also serve as the deposit address for other transactions. The visual model of this mode is shown below. The blue circle represents the deposit address, and the green circle represents the withdrawal address.
4. Inter-exchange transactions
The following figure shows the transaction model that occurred between Poloniex exchange and Binance exchange:
5. Common consumption transaction for withdrawal
Unconsumed transaction output (UTXO) is an important model that exists in cryptocurrencies such as Bitcoin, Litecoin, or Bitcoin Cash. From the perspective of the exchange, many transactions use unconsumed transaction output as the input for a particular transaction.
In the figure below, the small pink circle represents an unconsumed output, which is combined with the funds in the main transaction wallet to distribute the funds to the withdrawal address.
It should be noted that the other unconsumed transaction output (big pink circle) will be assigned to a different address.
As you can see,
machine learning
, and
advanced data visualization
combined explain the behavior of the exchange center of an interesting model. Sometimes, smart data visualization will stimulate the magic of human intuition to identify scenarios that are missed by machine learning models.
Centralized exchanges are still one of the black boxes of the crypto ecosystem, but
machine learning and data visualization technologies will help us uncover their secrets little by little.
Recommended reading:
There is a surprise by stamping "Read the original"
Is Lao Tie watching? ?