Due to the heavy desire from Venafi for the community to come together here, I thought I would share my experience with a total migration to Venafi as a Service from Trust Protection Platform, why we decided to make the move, and the good and bad things that occurred during the transition.Before I go much further, I do want to say that the Venafi team, and those of you I have worked closely with over the last year, are awesome. I am loud and opinionated, and youtake my loud opinionated self in stride. In all seriousness though, these are some of the best people from a vendor company I think I have worked with. I'd call you all out, and I will if you tell me too, but for now, you know who you are.I took over the team deploying and using TPP in October 2021. The team at that time had little leadership and was struggling to get it off the ground. TPP as it sits is a great tool, and if it were not for our infrastructure and the way our networks were designed, I would still be on it.The decision to migrate came down to three problems:30 millisecond latency TPP requirement to have a direct connection to Active Directory for SSO Single pane of glass for multiple security zones and global regions. Let me explain these two problems, and why they were detrimental to my organization's deployment:30 millisecond latency limitationTPP begins to break when network latency exceeds 30ms and can break entirely at 50ms.TPP was designed for datacenter use. It was never designed or intended for a global deployment with systems on every continent, where the organization was not paying for direct connections or cloud direct connects (pick your flavor of cloud and their direct connects). It was designed to handle certificate operations where bandwidth is very fast, locally.We use a lot of VPNs to get around the world. And VPNs add latency alot of latency. We havea use case where the application / dev team needed to be able to pull certificates programmatically across the internet and into their cloud environments, install the certificates in clusters immediately, on the fly, and into as many systems as any given customer requested. A VPN was established, code was written, testing began...and failed. 1 in 5 requested certs arrived. No matter what we did, we could not get more than 20% success rate. In troubleshooting the issues with the Venafi support team, we are told about the 30ms rule. I remember reading about this in the manuals, so I go back and re-read the section. Dang. They recommend getting hard lines to all of our datacenters and cloud providers globally.We got some quotes, and just to provide hard lines to our cloud environments was going to run us $12M a year.Um, that's not going to work.Active Directory and SSOWe are a major software company with our hands in lots of pots, including PCI and Financial Services, Hospitality, etc. To be honest, there is probably not a day that goes by that anyone doesn't touch our company services in one way or another.TPP was deployed in a global, cloud based, PCI zone. This was before I took over the project. The zone had its own AD, and we could provide self-service functionality to any user within that zone via SSO. Cool.But we were managing all of the certificates for the entire company, a lot of which is not PCI. We needed to connect to a different AD, one that is not PCI compliant, to provide SSO access and self-service to hundreds of additional users. TPP AD connectors only works with direct connection (I forget the exact terminology at the moment) to pull down AD users and groups. It did not support just-in-time provisioning.Can't make that work in a PCI zone without extending the zone, and we did not want to extend to this place. Working with Venafi support, they recommended that we deploy different TPP servers in the different security zones. But we wanted one pane of glass for global operations.TPP is an awesome tool. It is super mature; it has fewproblems. But we could not engineer our way out of these two problems built into TPP. It just couldn't be done and meet the rest of our requirements.Ultimately, Venafi recommended Venafi as a Service. We jumped at it.Venafi as a ServiceWithin a week, our VaaS instance was built, we were doing discovery scans of every IP and domain we owned, our discovered inventory was populating very quickly, and we were pleased.I read as much of the VaaS documentation as I could, to understand what we were about to do. To be honest, there isn't much to the VaaS documentation. It is nowhere near as complete as the TPP manuals. Venafi team - really need to improve the vaas docs. hint hint.I built the first vSatellites (required extra infrastructure!), configured the CertificateAuthority, built the first templates, created the first test application, and successfully issued testing certs. Nice. Easy.This is where we hit our first road bump.VaaS does not ingest keystores.It doesnt take JKS, PFX, or PEM files. Period.Hard stop. Was this in the docs? I dont think it was, or even is today.It will ingest public keys / certs only! And it will only accept one public key format, base64. You cannot upload other cert formats i.e p7b for example. This is sometimes frustrating, and I will cover this in another post.If anyone here does not know, to make VaaS do certificate issuance, you have to build a vSatellite (vSat). The vSat holds a private DEK that ensures your certs in the vaas cloud can only be decrypted and used by your team/company. This is actually a really great concept. And it works well, but only when you are creating a certificate in VaaS. It cannot receive a keystore file and encrypt the certificate during ingestion. (Yet. I am told this is coming. Im not trying to cover the conversations that I have had with Venafi, but supposedly this issue is going to be resolved).This was the first major roadblock for us and required us to step back and decide how we wanted to proceed. I didnt remember reading in any of the documents that VaaS could not ingest keystore files.Ultimately, we decided to migrate the public keys, get them into inventory, and we will renew them in VaaS as they come up for renewals. This was a big downside for us, because we wanted to shutdown TPP ASAP to remediate some risks we took to make automation work.Using the Venafi PowerShell module (love this by the way, thank you Greg for creating this), I was able to get the public certs downloaded from TPP in the correct format, and uploaded to VaaS.Phase one complete..