Installing Rubin Science Platform (RSP) on Local Systems

Support Rubin Science Platform
#1

Hello everyone,
I hope you’re doing well. I have a question regarding the Rubin Science Platform (RSP). My team is interested in evaluating the possibility of installing the RSP on our local systems here in Australia.
I tried to look for the source code or relevant installation documentation on the LSST GitHub page but couldn’t find anything specific. Before diving deeper, I wanted to reach out to the community for guidance.

  1. Is it possible to install the RSP on local systems outside of the LSST cloud environment?
  2. If so, could you please point me to the source code repository, installation instructions, or any documentation that could help us set it up?

Any insights, references, or advice on this matter would be greatly appreciated!
Thank you in advance for your help.

#2

Hi,

The CADC is in the process of installing the Rubin Science Portal (ie, just the Firefly interface) on our system. We’ve got it running for the public tables on our TAP services. But we haven’t got it running for proprietary data and we haven’t yet released an external visible instance. There may be an opportunity to work together on this.

Stephen

1 Like
#3

So you can and you may (and other institutions have) but whether you should is a more complicated question.

The Rubin Science Platform is what we call the user-facing side. From the point of running it, what you are talking about is running a phalanx cluster (phalanx.lsst.io) which is the platform that Rubin Services write on. There are install instructions however it is not turn-key (there is extensive reliance on services provided by IT infrastructure, the biggest being Kubernetes) and while it packages up the environment effectively it’s not turn-key and it doesn’t at this point save you from understanding how a system like this is put together. It also requires ongoing maintenance, as we are still evolving it and its services in non-backwards compatible ways.

Basically: Running a phalanx cluster (and hence the RSP) is a non-trivial commitment and has a fairly high devops technical floor. If you would like to talk more about whether this is a good value-for-time proposition for you feel free to DM me.

2 Likes
#4

Thank you both for your responses.

As Frossie mentioned, the key question for us right now is whether we "should " proceed with this or not. I will share this information with my colleagues to help them make an informed decision.