Over Concerns About Openness, OpenStack Founder Leaves Rackspace(readwriteweb.com)
readwriteweb.com
Over Concerns About Openness, OpenStack Founder Leaves Rackspace
http://www.readwriteweb.com/cloud/2011/03/questions-about-openstack---is.php
3 comments
This is a common story, engineer on project X does his best to convince management of X that they are making a mistake, they don't make the change, engineer leaves.
The variables in this particular tale are
* Engineer's ability to influence varies by factors {in|out} of their control.
* Management {listens | doesn't listen} and {ignores | decides} not to act on the suggestions.
* Engineer { leaves | is fired | stays and whines } about the whole affair.
Its one of the rhythms of work. Everyone gets to keep a score in their head of when they were right and when they were wrong. Its been said that if you're right 51% of the time you're ahead of the game. Personally I prefer to be right at least 61% of the time in ensure that being right isn't just a statistical fluke :-).
But listening to people's concerns, asking for them to explain their reasoning, and then acting based on reasoned debate rather than reputation, personality, or prejudice is key to getting to the best answer at a given time.
The variables in this particular tale are
* Engineer's ability to influence varies by factors {in|out} of their control.
* Management {listens | doesn't listen} and {ignores | decides} not to act on the suggestions.
* Engineer { leaves | is fired | stays and whines } about the whole affair.
Its one of the rhythms of work. Everyone gets to keep a score in their head of when they were right and when they were wrong. Its been said that if you're right 51% of the time you're ahead of the game. Personally I prefer to be right at least 61% of the time in ensure that being right isn't just a statistical fluke :-).
But listening to people's concerns, asking for them to explain their reasoning, and then acting based on reasoned debate rather than reputation, personality, or prejudice is key to getting to the best answer at a given time.
I wonder if they're a back story on this one...
How many of us could justify spending a lot of time on running our own object store rather than using S3? It makes sense that Rackspace is a lot more interested in it. I'd say that you'd need to be in the several terrabyte range before it makes sense to think about it. Clearly sites like Facebook, Flickr, and imgur want to run their own storage - they store a lot. But how many web firms do? I'm not saying it isn't an awesome project and I was excited when it came out. However, I'd have to be working on a site that stored a lot to justify building my own cloud storage cluster.
For what it's worth, it would be awesome for everyone if this engineer went on to create a Cloud Files and Cloud Storage competitor using Open Stack and worked with Rackspace without any grudge. Sometimes that's how openness happens. However, there tend to be fewer infrastructure vendors than web apps and so an infrastructure project like this can't be expected to garner the same attention as something like Apache or nginx which lots of people will be running rather than mostly a few infrastructure vendors. For most web operations, it doesn't make sense to pay a lot of money to build this infrastructure when you can get it for cheaper from Amazon, Rackspace, and others.
It might not be that Rackspace is being closed. It might just be that there are few operations that want to help develop these tools and most of them already have their own tools. Flickr and Facebook aren't going to drop their storage engine for OpenStack. Amazon isn't going to replace S3's software with OpenStack. SoftLayer does have their own cloud storage solution, but it isn't as widely used so there's a possibility there, if not a large one.
It's great that it's open source and if Rackspace keeps up the development and it's of high-quality it might start getting used by the next Facebook and Flickr style companies that need that type of storage. But it will happen slowly compared to the adoption of nginx, Django, MongoDB, etc.