We had a manager who thought that Cloudify could be used as a replacement for Horizon in OpenStack, but we found that Cloudify lacked the user interface or GUI for doing multitenancy and basic platform management tasks. 

Cloudify was really good at launching, for example, firewalls and configuring them and doing service chaining and rather advanced things like that, but it didn't meet the requirements for a basic platform management solution. It is something that seems to work better as a bolt-on or an augmented solution. It is a bit mis-marketed as a Cloud Management solution. It is not that. It is more of a service orchestration and automation tool. It is very good at doing that, but it fails to meet basic platform management requirements.

Once you have it running, you can't really do anything with it without writing code and scripts. It requires a full-time DevOps person to use it. We deployed a Palo Alto firewall with it. That's basically what the project was for us, and it worked flawlessly once we got it finished, but it took another 12 weeks to get all of the automation and everything else coded, tested, and working. There is certainly a place for this technology, but when we got rid of OpenStack and moved to VMware, we either had to go with the vRealize Automation Suite to do this kind of automation, or we had to find an alternative solution to manage the private cloud. So, we put Cloudify in, but we really couldn't find it useful for basic platform administration tasks.

We have solved many critical problems for our customers with Cloudify, including DevOps deployment, hybrid cloud management and VNF. These implementations have saved millions of dollars in operational and cloud management costs for our customers.

It's easy to create a cloud server, such as Linux with JBoss or WebSphere, which is good for continuous deployment.

Reduced the manual tasks during application release activities.

I think about the containers (Docker). We can quickly create a local cloud with Kubernetes, a global deployable format (container), good resource utilization, and open source. 

Product has given us the ability to catch scaling issues early on that many companies hit on with private clouds.

We’ve found application modeling being abstract blueprints fits the need for developers to implement their much needed bi-product from a simple template.

It has enabled us to create the DICE Deployment Service. It also simplified deployment of any other services (monitoring, continuous integration, etc.).