During a recent call with a potential business partner, a couple comments and questions helped me zero in on Tincup’s value.
“Are there really that many physical servers any more? Aren’t all the servers already in VMs? VMware has been around for a while now.”
“For customers that want to migrate to AWS, aren’t there already migration tools that make this easy?”
I’ve seen up-close why Tincup is necessary and explained Tincup’s core value proposition along these two dimensions.
- Physical servers are real
- Migrating to the cloud must be easy
Physical Servers Do Exist
This may not be widely known, but there are still today 10s (maybe 100s?) of thousands of physical servers in racks inside corporate DCs, in wiring closets, makeshift QA “labs”, under desks… all running some piece of an enterprise’s core business.
These servers often aren’t noticed (or at least remembered) until they fail. But they do fail, all the time. Expensive name-brand servers are usally under a third party service agreement and get repaired.
They’re running LOB processes, so something is done to put them back in service. When the issue is software-based, changes are made directly in the production environment and - surprise! - almost never get documented.
That one-off server just became a little more one-off. This is common and represents a substantial risk that’s difficult to manage.
The business processes running on these servers are prime candidates for migration to the cloud, but the LOB itself can’t be disturbed during the refactoring that must occur before the legacy services run in the cloud.
The solution to this problem is to use Tincup to decouple the hardware risk from the refactoring. Cloning the servers in Tincup gives the team a 100% fidelity copy of the still running production server. Refactoring can begin in the clone without even touching the running production systems.
Migrating Must be Easy
Yes, AWS provides a migration tool that assists in migrating VMware machine images to EC2. While this is great for some enterprises, a large number - majority? we don’t know yet - won’t use these tools.
Why? They’re not easy to use, relatively speaking. Skills + priorities often aren’t aligned with the requirements to use these tools. This leads me into a critical point.
If you want to make a dent in the enterprise, create a tool that makes using critical technology easier.
containers->k8s; k8s->heptio; images->tincup
Tincup is point + click simple. It generates a 100% fidelity clone of a boot disk + transforms that to an industry standard image format which is then used to generate a bootable machine image that can be shared with an AWS account.
Now’s that’s easy.