Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Estimate Hardware requirements, CPU, RAM and NVMe disks #220

Open
Tracked by #221
a-saksena opened this issue Sep 30, 2024 · 0 comments
Open
Tracked by #221

Estimate Hardware requirements, CPU, RAM and NVMe disks #220

a-saksena opened this issue Sep 30, 2024 · 0 comments

Comments

@a-saksena
Copy link
Contributor

a-saksena commented Sep 30, 2024

Known questions:

  • Main disk requirements, RAID/non-RAID options, maybe ZFS or btrfs?
    • How much disk space gets us how many years storage?
  • What is the hash rate, can we use GPU or NPU resources, and what accelerators are available?
  • What is the networking required, and what are the options?
  • How much RAM do we require, what are the primary drivers of RAM usage?
  • How much parallelism do we require, how many cores?
  • Can we offload some functions to separate machines, possibly ephemeral?
  • Can we use "typical" cost reduction approaches?
    • Can we reduce cost with several small/cheap machines?
    • Can we lower cost with networked storage for the large-volume historical data?
    • Can we use separate "worker" processes (K8S containers) to handle large/long processes like reconnect or state snapshots?
    • Should we be using ZFS, btrfs, LVM, etc...?
    • Do we need RAID (5/6) or is multiple-machine redundancy enough/better?
  • What uptime requirements do we have?
  • What is the impact of 1/3/6/12 simultaneous block node failures?
  • What level of "spare" capacity do we need to avoid failure amplification?
  • What is the disk read/write rate required for "recent" blocks, live state, state snapshots, "old" blocks, etc...?
  • What data volume is required for "fast" data versus "slow" data?
  • Do we need/want any level of CXL 2.0 support?
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant