Ohh crap. First Enterprise Build - I have Qs

Hello guys,
As the title says, I’ll be building a system using enterprise hardware and it will be my first. I do have some questions that will determine what hardware I go with. Also, this project for me is above my current IT knowledge, though I I’m confident I can succeed.

I want to thank you in advance for bestowing your knowledge to me.

Project rundown:

The system’s main purpose will be a Full Binance Smart Chain Node (crypto node). This node will have to be synced 100% of the time. The reason for this is, I am developing an app that tracks targeted wallets and their transactions in real time and displays them to a few for analysis.

Communication speed between the app and node is of the utmost importance. For now, unfortunately the app will be windows based vs the node running on Linux.

60% of the time I will be using the app remotely, the other percentage I would like to have the app running on the same system and even be able to VPN in sometimes.

Budget $8000 USD

Hardware:
CPU: AMD Epyc 7542 32core 2.9ghz
RAM: 128 GB
Mobo: Asrock Rack ROMED8-2T
M.2: 2 x Samsung 983 Dct 1.9tb (Raid 5 config in future)

Questions:

  1. Considering the above, would using VM be efficient? Or connecting 2 systems be a better option for communication and speed?

  2. Can I run just a VM for windows? (Positive I can)

  3. If going VM to VM, networking is doable on same host?

  4. Fast, low latency remote management / VPN recommendations?

Again, thank you.

1 Like

Does the host system matter?

If the host is MS Windows, then why not run the Linux stuff in WSL2.
If the Host is GNU/Linux, then run MS Windows in a VM and you can run any of the additional Linux stuff in a LXC instead of a full blown VM.
The number of VMs depends on how much isolation you need between systems.

Need clarification here. If you are planning to access this through the Internet, then the latency and speed will all depend on the ISPs at both ends and the paths used to get to them. Basically you will have to deal with the worst path traveled and develop with that in mind; you will not be cable to control this. You can mitigate that by having a server that these the client and server talk to, but that adds complexity.

2 Likes

Sounds like an interesting application.
But to boil it down you just want something to effectively track one (or many?) blockchains and effectively monitor wallets.

And then it outputs data either on a web portal or email or other message service?

I am not into cyrpto much, But I would probably say it might be worth considering prototyping what your software does in a cloud instance. Then you’ll get a gauge of what grunt you need. So for a more long term gain, get the hardware spending the money on where it counts. This will also let you know how to architecture the software side such as VMs better. But docker seems to be the way things run to simplify yet give better flexibility.

Limit your resources and work up as to not drain your wallet too fast.

What kind of timescales are you thinking of developing this in?

2 Likes

This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.