若干中国区eos节点的硬件配置

in #eos7 years ago (edited)

节点硬件配置

1. EosONO :

Estimate of Technical Specifications and Total Expenditure of Resources by June 3, 2018:
Amazon EC2
AWS m4.4xlarge
CPU: 16 Core
RAM: 64GB
SSD: 512GB

Total Cost: USD $817 per month

Estimated scaling plan for hardware after June 3, 2018:
Amazon EC2
AWS x1e.32xlarge
CPU: 128 Core
RAM: 3904G
SSD: 2 x 1900GB (The disk capacity increases as required)
Network: 25 Gbps
Nodes: Total 2 Nodes (1 livenet, 1 standby)

2. Eos Shenzhen:

1、Primary Location
Block producers Primary:
Dual Xeon, 128 Gb ram, 512 Gb SSD

Full node:
Dual Xeon, 64 Gb ram , 512 Gb SSD

Storage (ipfs):
Dual Xeon, 64 Gb ram , 128 Gb SSD, 40T HardDisk, 7.2k rpm
In the Internet Data Center. with 1gps bandwidth, and it can board enough if you need, for example it need to board to 5gps even more, we can make it.

2、Secondary Location
Block producers backup
1:Dual Xeon, 128 Gb ram , 512 Gb SSD Full node & Storage
(ipfs) backup
1:Dual Xeon, 64 Gb ram, 128 Gb SSD, 40T HardDisk, 7.2k rpm
In the Internet Data Center. with 1gps bandwidth, and it can board enough if you need,
for example it need to board to 5gps even more, we can make it.

Total Expenditure
The budget is sufficient, we have other income from various investment.

Estimated scaling plan for hardware after June 3, 2018
Plan to use:
1, the most security Internet Data Center, three adresses at least
2, Higher configured hardware,for example 2TB RAM
3, Enough bandwidth support, the largest is 10gps

3. ChainPool

Estimate of total expenditure for resources by June 3, 2018
CPU: Intel(R) Xeon(R) E5 2697 v2 2.7GHz 12 Cores
RAM: 128 GB DDR3 RECC
Storage: 1 internal 1 TB PCIe SSD, 5 external 2.0 TB Intel P3700 PCIe
Network: 1 Gbps

Estimated scaling plan for hardware after June 3, 2018
Initially, we are planning to have a primary and a backup machine as main producing units and are looking for more effective ways of power supplies to minimize the loses. Our hardware spec listed below is subject to change based on Block.One and feedback from the community:

CPU: Intel Xeon E7 8880 v3 (18core) * 4 = 72 cores
RAM: 1- 4 TB
Storage: 100+ TB
Network: 25 Gbps
Nodes: 1 livenet and 1 standby

4. Whateos

Technical specification by June 3 2018:
BP nodes:
3 x Alicloud ECS (Local SSD Type i2, ecs.i2.4xlarge)

1 x SLB:

1 x VPN, 1 x WAF, 1 x DDOS

*our initial MVP setup does not expect a high user volume before or soon after launch of EOS main net. The additional servers will be provisioned close to activation date to compensate the increase in volume.

Technical specification after June 3 2018:
Multiple Local SSD Type i2,ecs.i2.8xlarge across different regions

SLB:

Additional DDOS and Kubernetes container specifications will be selected as the situation requires.
Further local server bare-metals set up will mimic the same set up in cloud, however, specific price will vary based on charges of local supplier and local data center providers.

5. 91EOS

• BP Servers (1 livenet, 1 standby): 16 Cores; 256G RAM; 1T SSD Hard Drive
• API Server: 16 Cores; 256G RAM; 1T SSD Hard Drive

6. Helloeos

In testnet and early stages when block is small and load pressure is still low, we adopt the cloud based server approach:
• BP Servers (1 livenet, 1 standby): 16 Cores; 256G RAM; 1T SSD Hard Drive
• API Server: 16 Cores; 256G RAM; 1T SSD Hard Drive

7. Oraclechain

Now OracleChain is the testnet block producer node, and we adopt the DigitalOcean cloud machine with the server approach like:
12core CPU(Intel Xeon ProcessorE5-2650 v4)/48GB DDR4/320G SSD
When the EOSIO mainnet is running, we will adopt the cloud machine with the follow basic configuration like:
(1 host, 1 standby)x1.32xlarge/64 core/1024GiB Meory/10TBSSD (We use the AWS cloud server as an example. Actually, other cloud host services or IDC servers may also be used.)

8. Eos Gravity

Master Server:
PowerEdge R730xd
4 CPUs Intel® Xeon®
128G * 2 (2400 MHz) 内存
300G * 2(系统盘) 7.2k rpm
512G * 1(数据盘) SSD
双电源

Slave Server:
PowerEdge R730xd
4 CPUs Intel® Xeon®
128G * 2 (2400 MHz) 内存
300G * 2(系统盘) 7.2k rpm
512G * 1(数据盘) SSD
双电源

Switch:
H3C S5130 48口 * 2
WAF(Web Application Firewall 应用防火墙):
WebRAY Ray-WAF-1006P(需要购买license)

9. EOS.CYBEX

目前的技术规格和支出在6月3号之前

Server Specification

1 x Cloud Node(Microsoft Azure D13 V2)
CPU: 8 vCPU Cores
RAM: 56GB
Network: 1000Mbps
Storage: Local SSD 400GB, Managed OS Disk 1TB
Estimated Monthly Cost: USD $711
6月3号之后的硬件扩充计划
我们计划在两个不同的云服务供货商(IBM和微软云)上设置服务器,以保证系统供货商,云计算,网络连接都有足够的容错性 。

主要服务器

1 x Bare Metal Node (IBM Softlayer BI.S1.H1000)
CPU: Quad Intel Xeon E7-4890 v2 (60 Cores, 2.80 GHz)
RAM: 1TB RAM
Network: 10Gbps
Storage: 4 x 800GB SSD (Expandable to 17 x 800GB SSD)
Estimated Monthly Cost: USD $6,058 / month
备用服务器

1 x Cloud Node (Microsoft Azure E64 V3)
CPU: 64 vCPU Cores
RAM: 432GB
Network - 30Gbps
Storage: Local SSD 1.6TB, Managed OS Disk 2TB
Estimated Monthly Cost: $4,688
一旦今年谷歌的服务器系统在香港可行,我们可能增加或转移到谷歌的服务器上,用来提供正常运行外的容错。