Browse Prior Art Database

Method and System for a Scalable Hierarchical Concurrent Update Tree with an Optimized Response Time for Processing Transactions

IP.com Disclosure Number: IPCOM000249452D
Publication Date: 2017-Feb-27
Document File: 2 page(s) / 33K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method and system (TreeLedger) is disclosed that implements the BlockChain interface but internally provides several queues organized as a tree into which a transaction can be inserted for processing. The method and system includes multiple load balancers in front that arbitrate the incoming transactions using lock-free queue status.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

1

Method and System for a Scalable Hierarchical Concurrent Update Tree with an Optimized Response Time for Processing Transactions

Current BlockChain implementations are limited to a small number of transactions per second (TPS) due to the consensus protocol that puts network latencies into the critical path since all transactions in a channel are serialized. This leads to BlockChain systems that do not scale well if implemented as a distributed system. Also, the use of a single master implementation removes the critical value that a distributed approach offers. For markets such as, but not limited to, the credit card business and high volume trading it is essential to reach high TPS.

Disclosed is a method and system (TreeLedger) that implements the BlockChain interface but internally provides several queues organized as a tree into which a transaction can be inserted for processing. The method and system includes multiple load balancers in front that arbitrate the incoming transactions using lock-free queue status. Further, unique monotonous timestamps are used to detect and ensure a sequentially consistent ordering of the dependent transactions. It is not required but possible to order the independent transactions by timestamp to achieve a well-defined global ordering of all transactions across all master nodes. The method and system is optimized for a particular usage scenario using the tuning parameters such as, but not limited to, tree topology (for example, depth) and timeout settings.

The tuning parameters of the TreeLedger include parameters such as, but not limited to, m (fan-out at the master), n (fan-out for intermediate nodes), number of levels (tradeoff larger blocks versus more blocks at the master), timeout per level and queue size per level.

The following are the properties of the method and system disclosed herein. The method and...