Predictive normalization for blockchain resource management
AssigneeCelligence International LLC
InventorsPavan Agarwal, Gabriel Albors Sanchez, Jonathan Ortiz Rivera, Thomas Smith
Life of this patent
Estimating the term…
Abstract
Systems and methods are provided for leveraging a normalization algorithm to optimize resource allocation dynamically. The algorithm detects deviations that indicate potential increases in demand. This predictive capability allows the system to scale resources proactively, adding signer nodes when necessary to handle increased transaction volumes. By normalizing transaction data and using it to predict and respond to changes in system demand, blockchain networks can maintain optimal performance and efficiency, preventing service degradation and ensuring scalability in a cost-effective manner.
Claims14
A computer-implemented method for scaling nodes in a blockchain network, the method comprising:
- collecting historical node data comprising historical unconfirmed transaction counts and associated timestamps for each node;
- applying a normalizing algorithm to the collected historical node data;
- training a selected machine learning algorithm on the normalized collected node data;
- executing the trained machine learning algorithm to determine a number of unconfirmed transactions to occur within a time period within;
- and
- establishing a threshold based on the determined number of unconfirmed transactions to occur within the time period;
- continuously monitoring a current count of unconfirmed transactions for each node by applying the established threshold;
- upon determining that the current count of unconfirmed transactions for each node exceed the established threshold, initiating a vote for adding a new node;
- and
- upon the vote passing, adding the new node into the network, wherein the new node is a voter node.
A system for scaling nodes in a blockchain network, the system comprising:
- one or more computing processors;
- and
- a machine-readable storage medium storing instructions that, when executed by the one or more processors, cause the system to:
- collect historical node data comprising historical unconfirmed transaction counts and associated timestamps for each node;
- apply a normalizing algorithm to the collected historical node data;
- train a selected machine learning algorithm on the normalized collected node data;
- execute the trained machine learning algorithm to determine a number of unconfirmed transactions to occur within a time period within;
- and
- establish a threshold based on the determined number of unconfirmed transactions to occur within the time period;
- continuously monitor a current count of unconfirmed transactions for each node by applying the established threshold;
- upon determining that the current count of unconfirmed transactions for each node exceeds the established threshold, initiate a vote for adding a new node;
- and
- upon the vote passing, add the new node into the network, wherein the new node is a voter node.
Dependencies are inferred from the claim text.