Hacker News new | past | comments | ask | show | jobs | submit login

A couple of points where we differ:

1. Deployment - basically, boot a seed node, then turn on machines. New nodes automatically install and come online, and are as easy to remove. This eases management when you're dealing with hundreds or thousands of machines.

2. Permissions - we've got a much finer-grained permissions system which comes in useful if you're deploying a massive cloud for use across, for example, a multi-site and multi-department organization.

3. Self-healing - we've got a lot of monitoring stuff which watches running services and migrates them around the cloud as things (inevitably) break or become unavailable.

4. Layer 2 networking - we're providing virtual layer 2 networks amongst instances, which allows dynamic creation and re-organization.

5. Multi-tenancy - there's a full multi-customer, multi-user model in the system, which allows you to easily re-sell capacity.

6. Federation - using the same user and permission framework, you can pull up instances either in the local private cloud, other people's Nimbula clouds, or in public clouds like EC2 or Rackspace (or, I guess, even accessible Ubuntu clouds).

Basically, UEC is a good choice for running a local EC2 clone, but falls down a bit when you require very large clusters or strong security. We've had the opportunity to rethink and design our own system, rather than clone an existing one, which allows us to offer a greater range of possibilities.




It will be difficult to make inroads against HP Server Automation, but I am sure people would love to see you succeed.

That being said, this is very much a solution services business. You will need more than core talent. A lot more.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: