This paper

Currently, there is a contradiction between the expansion of a blockchain network itself and the massive growth of local applications under the system, resulting in intensified competition for system resources, which has been reflected in the resource congestion (RAM and CPU) in the CURRENT EOS network. Not only EOS, but also blockchain networks such as Ethereum and Bitcoin.

The solution is Polkadot, a scalable heterogeneous multi-chain system that aims to interoperate with isolated blockchain networks. Polkadot has high scalability, avoids hard forks, and can be continuously upgraded.

IPSE constructs parallel chains (Parachain) based on Substrate architecture and uses the main key structure of Polkadot system architecture, the transfer bridge (BRIDGE), for cross-chain communication. To ensure state consistency between IPSE parallel chain and Polkadot relay chain, a hybrid consensus mechanism is used, namely BABE and GRANDPA, respectively.

BABE consensus algorithm

BABE (Blind Assignment for Blockchain Extension) is a block generation mechanism between verifying human nodes and deciding who is generating a new block. The BABE algorithm allocates block production slots to verifiers based on collateral and using Polkadot random cycles.

Polkadot verifiers will participate in each draw to determine whether they are production block candidates for that time slot. A time slot is an interstitial unit of time and is 6 seconds in length. Because of this randomness mechanism, multiple verifiers may be candidates for the same time slot. The rest of the time, the timeslot may be empty resulting in inconsistent block times.

Multiple verifiers in the same timeslot: When multiple verifiers become block production candidates in the same timeslot, all blocks will be generated and broadcast to the network. This scenario, though, would be a race. Which verifier wins first depends on the network topology and latency until the fork is finalized and truncated.

No Verifier in time slot: When no verifier qualifies low enough in a random draw to produce blocks, time slots are created without blocks. However, algorithms are chosen to avoid this, keeping a verifier always generating blocks.

BABE also has to work with GRANDPA, who is responsible for generating blocks, and GRANDPA is responsible for achieving consensus between blockchains (such as parallel chains) to ensure the block’s final consistency.

Note: The BABE algorithm is still a work in progress, and Handan Kilinc Alper at the Web3 Foundation has published a new BABE documentation that explains how it works.

Research. The web3. Foundation/en/latest/p…

IPSE search ipse. IO