Survey deployment tools

[mhchia]

What is wrong?

Previously I used fabric script to deploy and synchronize the code across servers and spin off the nodes. However, the script is a little bit messy for now and fabric is not that powerful. (e.g. for parallel commands you can only execute the same command among all servers). So I think it is a good timing to survey other options.

How can it be fixed

Survey both local and cross-servers deployment

• Use containers(related to https://github.com/ethresearch/sharding-p2p-poc/issues/13 ?)
• ansible for smaller scale
• kubernates for larger scale
  • example: ipfs with kubernates
• Survey terraform for managing EC2 instances

[jrhea]

I have a couple of questions on the requirements:

• how many nodes do u want to simulate?
• do you want them to span different regions? Or do u want to just simulate latency?

[mhchia]

@jrhea

1. 10,000 nodes should be sufficient
2. Do you mean testing? If so, yes they should be spanned in different regions.

[ChihChengLiang]

Addressed in #73 and #77. Chose the combination of Terraform and Ansible