What is our primary use case?
We are working on two different deployments, one based on WSL, which is like Windows WSL, and the other one is Ubuntu. We have a local setup deployment that basically serves as a local AI inference agent based on NVIDIA GPUs, and we use Kubernetes with Docker.
What is most valuable?
First of all, since it is cross-OS, it is very easy for us to do small deployments on edge computers that are using Windows with Ubuntu, and then do the full production on Linux local servers. We use this frequently, and they have made a lot of progress lately with their cross-integrations and WSL integrations and Kubernetes.
A second thing that we really appreciate is that it is very easy for people to understand because it reminds them of the Git workflow, so all the image management is so convenient that people feel they are working with Git. Here they are not managing code; they are managing images.
Since we are doing a cross-compile, we want to decouple it and basically have a constant image for our environment. We are building OS inside OS, and we are using this feature. It allows us to create an independent cross-compile image that is basically decoupled from the host.
For us it is a small tool; it is not a big investment, so we are not tracking the ROI of the product directly, but I can tell you that it gives a lot of useful features and overall the ROI is worth it. Especially, now they have this MCP integration and all management of all MCP servers is in this new feature, and it basically allows you to deploy MCP servers securely because each MCP server is running in a separate Docker. In combination with all those features, the ROI is good.
What needs improvement?
I think the licensing is still confusing. Since it is so easy to install, the licensing remains confusing regarding what is for enterprise and what is for personal use. Basically, you can install it, but then you get a message that this is personal mode, and people are afraid whether they stand with the license or not. The licensing is a little bit confusing.
A second thing is that there is this Docker Hub, and I don't know why, but people say that they have issues with the images that they download there. Personally, we don't really use Docker Hub.
For how long have I used the solution?
We have been using this for many years, maybe five years.
What do I think about the stability of the solution?
In combination with Kubernetes, we don't see any limitations. There is now a composer, so sometimes people use composer, and many now use image composer images.
What do I think about the scalability of the solution?
In combination with Kubernetes, we don't see any limitations.
How are customer service and support?
I don't have something special to report, but I can tell you that it passed our internal CISO review. We are working with top secret data, and we have a Common Criteria regulation that we need to comply with. From a security specification standpoint, it was accepted.
Which solution did I use previously and why did I switch?
I am not sure of its competitors. I switched from working at Intel to a new company. At Intel, we had another solution; I don't remember the exact name, but it was more legacy, and everybody wanted to switch to Docker. It was something parallel to Docker, but eventually it was considered legacy.
How was the initial setup?
It is very easy; it is one of the easiest setups.
What was our ROI?
For us it is a small tool; it is not a big investment, so we are not tracking the ROI of the product directly, but I can tell you that it gives a lot of useful features and overall the ROI is worth it. Especially, now they have this MCP integration and all management of all MCP servers is in this new feature, and it basically allows you to deploy MCP servers securely because each MCP server is running in a separate Docker. In combination with all those features, the ROI is good.
What's my experience with pricing, setup cost, and licensing?
I am not paying for it, but I know that there was no issue. I don't remember the exact number, but comparing it to other costs that we have in other tools and especially in AI, it is reasonable. It is not something that is preventing the purchase.
Which other solutions did I evaluate?
At Intel, we had another solution; I don't remember the exact name, but it was more legacy, and everybody wanted to switch to Docker.
What other advice do I have?
We have a very high bar because we are developing a security product, so it is the highest bar that can be.
Basically, we have here two types of environments. One environment is top secret data that cannot leave a special network that is only inside the organization; we cannot use any cloud provider for it, so all this setup is running locally. We have the same images from an infrastructure perspective that is handling other projects that do not have such secure data; their data is less secure. It is still secure, but it is at a level that allows us to use cloud. It is not external cloud; it can be AWS or Azure, but still, it is allowed. The fact that we are using this hybrid environment means we are reusing the same infrastructure for different data type classifications.
We never reached the point that we need to search in community forums for help. My overall rating for this product is 9 out of 10.