Node Device

Node device In order to understand the position and importance of the index node in the node network, let’s take a look at the HyperGraph business data processing flowchart.

In this picture, from the middle to the lower right corner, the HyperGraph network analyzes smart contract events, processes and stores data, which is the work that the index node must complete. So the index node is the core of the core. Since indexing and storage are operations that require both a large amount of calculation and a large amount of storage, the index node needs to have better hardware calculation and storage resources, and install corresponding software facilities to index data and provide query services.

Depending on the amount of service, different hardware devices can be used for the hardware. The following is divided into three scales to introduce separately.

A. Small nodes

Small nodes are suitable for individual developers or small business teams to participate in the construction as amateur professions. Provides an index of less than 20 subgraphs, or the amount of data per block is in the order of <10 transaction processing (using 3s as the block generation time), and does not need to build a blockchain RPC archive node, but use the core developer to provide For free RPC resources, the recommended hardware configuration is as follows:

CPU

RAM

Hard disk

Description

8-core

16G

200G (SSD)

The community provides free RPC resources

The cost of this model on Alibaba Cloud is as follows:

But because HyperGraph nodes can get preferential prices for collective purchases, the final result is as follows:

B. Medium-sized nodes

Medium-sized nodes, suitable for medium-sized application application teams or medium-sized application teams contribute their own nodes to participate in the construction, provide an index of <50 subgraphs, or data volume within 100 transactions per block, 200 responses per second request. Medium-sized nodes can build their own blockchain RPC nodes, or they can use RPC resources provided by core developers, but additional HGT payments must be added for RPC resources.

CPU

RAM

Hard disk

Description

16 core

32G

500G (SSD)

The community can optionally provide RPC resources

The reference cost of this equipment discount is as follows:

C. Large nodes

Large nodes, suitable for large teams or professional partners to participate in node construction. It can store more than 100 subgraph indexes. The amount of data that can be indexed in each block is on the order of 200 transactions. It can provide more than 1000 query requests per second. Large nodes need to build their own RPC archive nodes and support index nodes. Vertical segmentation.

RPC node

CPU

RAM

Hard disk

Description

16 core

32G

6000G (SSD)

Need to support the space expansion at any time, can be opened to third parties to use

The cost of RPC nodes is relatively expensive. The main cost is in storage space. The reference cost is as follows:

Index node

CPU

RAM

Hard disk

Description

32 cores

64G

1000G (SSD)

Note: The above hardware prices are all reference prices. The actual price is based on the actual transaction. The actual transaction is the actual cost without other additional expenses.

In addition to index nodes, the software can also support splitting. The back-end software facilities can be split and deployed.

In terms of software, large nodes require specific environments such as the following software construction, compilation, and operation, as well as the basic deployment and maintenance capabilities of the corresponding environment.

  1. Linux operating system

  2. Docker container or Rust language compilation environment

  3. Nodejs development and compilation environment

  4. Postgres database

  5. IPFS

  6. HyperGraph node core program

  7. HyperGraph inode program

The support required for specific services and actual processes is implemented by core developers with corresponding help, programs, and scripts.

Last updated