I was previously CTO in SaaS company for 24 years.

It sounds like they are running a single instance for each customer - and wanting to move to multitenancy model where the same instance can serve multiple customers.

The former is how software used to be developed - and lends itself to running in a customer data centre or customer cloud account.

The latter is more complicated to develop from security perspective (to keep data logically separated) but a lot cheaper to run as it uses shared resources.

These days nearly all SaaS products are architected to be Multi-tenant

However as you point out @edgescape government and other categories such as financial services may demand a dedicated instance if the data is sensitive.

But they can still save costs by providing that option where required and a cheaper shared multi-tenant option elsewhere.

Also they could potentially group say all Australian government customers on one multi-tenant instance if that was acceptable

If they are redeveloping the code then as indicated:

• They can have a platform that supports both multi-tenant and single instance on same code base
• They can swap out the database to a cheap or free open source product (Im assuming the Microsoft licensing is the database)
• They can migrate to free/open source Linux from Microsoft (and reduce licensing costs)

I would have thought its a 1 (to 2 year max) project to covert over for products like Prophecy have.

The other comment is that they are recognising they need to invest to bring down hosting/cloud costs.

This is very common learning once you move into the cloud. In the beginning you follow 'best-practice' from cloud providers. Then you realise that leads to a lot of high costs and waste of resources. So its very common to then invest into lowering costs - which often requires changing architectures and moving to open source database systems and linux.

@shearman

Thanks for the detailed response

I am more on the sys support and arch side so I can understand where you are coming from at a high level.

Definitely there would be a separate pricing for dedicated - stupid not to if Prophecy was not doing that.

I don't know about moving to open source model as some are Microsoft only and don't want to mix with Linux.

There's lots of ways in approaching what prophecy is trying to do. I understand OCI has some bare metal and dedicated architecture offerings that can be used as well.

I'm thinking at best case they got the orchestration wrong like their deployments were provisioning more powerful computes or too many. Or they didn't have scheduled shutdowns of Non prod systems. At worst case there is something underlying in the architecture that we don't know about.

Or maybe they found a better way of redoing the Emite offering in a shared architecture as you said? I have some other thoughts on this but will keep it to myself for now because it is my personal opinion only.

Either way they have now acknowledged they are out of their depth and that is the risk I highlighted earlier about recruiting the best. I agree it will take about 1 or 2 years but the report has no time frame.

@shearman "This is very common learning once you move into the cloud. In the beginning you follow 'best-practice' from cloud providers. Then you realise that leads to a lot of high costs and waste of resources. So its very common to then invest into lowering costs - which often requires changing architectures and moving to open source database systems and linux."

I suspect this is bang on. PRO shifted their internal development to Oracle cloud a few years ago, and Oracle used them as a local case study for the efficiency benefits that would come:

https://www.oracle.com/au/customers/prophecy/

https://www.oracle.com/customers/prophecy-international/

A few years down the track it may be the case paying for the bells and whistles is unnecessary and the efficiency benefits don't offset. I will seek more clarification from management on what wrong and the expected budget to fix the issue (+50% because it's a large scale IT project), but hopefully they can avoid disruption to customers and growth.