Processing vs. Sharing WebODM on Cloud

WebODM
#1

Hi all,

I am starting to survey and process large areas (~1000 acres, 3000 images) and, as a result, moving away from Pix4D like I was using before. After doing some research, I have hired a developer to deploy WebODM through AWS on a r5.2xlarge spot instance. I know that this is the most cost-effective way to create the orthomosaics I need, but also that to keep the price down I will be turning the server on/off as needed. What is the best way to share the outputs in a way that my clients will be able to view them from their web browser even if the server is turned off? Is anyone doing something like this and can explain their process? Thanks in advance for the help!

1 Like
#2

Take a look at:

But the basic principle beyond these very clever projects is to separate your WebODM display node from your processing nodes.

1 Like
#3

Thank you for the quick response, I have to admit I’m completely out of my element here with pretty much no coding background. The guy I’m paying to set this up is an AWS Certified DevOps Engineer but don’t think he’s ever worked with WebODM. Will he be able to separate the WebODM display node from the processing nodes if I tell him that? Or should I look for someone who has more experience with WebODM?

1 Like
#4

Also, I know I need the r5.2xlarge spot instance for processing, but can I get away with a less expensive server for the display?

1 Like
#5

If you’re not going to let your clients do things like generate contours and elevation maps, sure, scale down a few.

If you are going to let them use those tools, they’re very processor-intensive and single-threaded at present.

#6

I would say only orthomosaics and elevation maps are important to me, no 3D modeling or contours really. How much RAM of a server should I try to get to allow clients to view orthomosaic and DTM/DSM?

1 Like
#7

Also, for the more powerful server I am using to process, only the NodesODM software needs to be deployed on this, correct?

1 Like
#8

Just view?

I’ve viewed large datasets from my phone and my netbook (both 4GB RAM), so I don’t think you need to go too big… But scale that horizontally for how many concurrent users you expect, etc.

#9

Yes just viewing and maybe some basic measurements, and I don’t expect more than 1-2 viewers at a time.

1 Like
#10

@Saijin_Naib, do I only need to deploy NodesODM on the server I want to use for processing?

1 Like
#11

That should be all, yes. And Docker/docker-compose, but yeah :slight_smile:

#12

Awesome! Thanks so much for all your help!

1 Like
#13

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.