AWS And OpenStack: An Interesting Contrast Worth Considering During Cloud Selection

 Having attended the OpenStack Design Summit this week and at the same time fielding calls from Forrester clients affected by the Amazon Web Services (AWS) outage, an interesting contrast in approaches bore out. You could boil it down to closed versus open but there’s more to this contrast that should be part of your consideration when selecting your Infrastructure as a Service (IaaS) providers.

The obvious comparison is that AWS’ architecture and operational procedures are very much their own and few outside the company know how it works. Not even close partners like RightScale or those behind the open source derivative Eucalyptus know it well enough to do more than deduce what happened based on their experience and what they could observe. OpenStack, on the other hand, is fully open source so if you want to know how it works you can download the code. At the Design Summit here in Santa Clara, Calif. this week, developers and infrastructure & operations professionals had ample opportunity to dig into the design and suggest and submit changes right there. And there were plenty of conversations this week about how CloudFiles and other storage services worked and how to ensure an AWS Elastic Block Store (EBS) mirror storm could be avoided.

Having this knowledge about the AWS architecture might have helped many Elastic Compute Cloud (EC2) customers work around the problem or even potentially detect it and act preemptively. Certainly all enterprise I&O customers should have employed disaster recovery and HA best practices with their cloud applications so any unforeseen issue could be dealt with. And that would be true no matter what cloud you choose.

Another stark difference is in the ecosystems of these cloud technologies. Amazon’s massive ecosystem of management software, SaaS, Platform as a Service (PaaS) and independent software vendor (ISV) partners are all just customers of AWS. They aren’t contributors, nor are they I&O peers. OpenStack’s ecosystem includes many of the above but also service providers, government agencies and enterprises who are implementing OpenStack as well. Which means that if you choose to implement OpenStack for your private cloud or you are a hoster implementing it for your IaaS, there is an I&O peer network you can call on in times of crisis who have intimate knowledge of the software and can help you diagnose the problem and even potentially supply a patch to reverse the problem.

You could argue that there is similar value to deploying VMware vCloud Director as I&O peers would be able to help as fellow administrators, but everyone would have to fall back to VMware to resolve the issue if deep knowledge or code changes were required.

The picture painted here is certainly not Nirvana. Open source projects are starting points, and commercial implementations often intermix this code with commercial code that mirrors more the VMware model. Plus every implementation will differ based on hardware used, components deployed and scale. But in all cases, the greater I&O community can be called upon to help in times of crisis and that’s worth considering when choosing your private cloud platform or public cloud provider.  

Has the AWS outage last week changed your opinion about using cloud computing solutions? If so how? Share your thoughts and feelings in the Forrester I&O community.

Comments

Value in masking complexity

There has always been value in masking or managing complexity. Cars have come full circle from needing to hire a mechanic just to own one (100 years ago) to anyone being able to tinker/repair and now back to being so complex that you have a hire a mechanic again. I miss the tinkering on cars era. I don't have the time or inclination to tinker on my technology.

Apple has consistently proved there is a population willing to pay for that convenience with Mac vs DOS and again with iOS vs Android. AWS may have taken the masked/managed complexity approach by design or happenstance. It's been profitable and may well continue for some time to come. The key is it all has to work and work VERY well. AWS' face plant wasn't terminal nor will it be. I expect they will continue to appeal to the mindset that wants all that magic dealt with by the people who know - so long as "architecting for failure" is well understood and implemented on the customer side.

OpenStack will continue to appeal to people who are unwilling to trust others to manage that complexity or those that are just bit-twiddlers by nature.

Both models have pluses and minuses but they both work. Both can be cost/benefit justified depending on the buyer persona.

True but hosters collaborating is the unique value

Well said, Tim. And very true. However, even those who want to trust their cloud provider would take solace in knowing that a community of cloud providers can work together in the background to help each other. Sure, competition might impede that but we've seen the security community collaborate against hackers and no one wants an HA issue to affect a peer than could later affect you too.

Third paradigm forming

James,

Indeed. Perhaps a third paradigm is forming from this. I hate the word 'hybrid' but perhaps 'mixed' comes closer. There is either Mac or PC, there is either iOS or Android. There is no mixing of the two whereas you can combine elements of AWS and OpenStack several ways based on your needs or risk profile. There will be tradeoffs in elasticity, complexity, etc. but it is possible.

I think it's cool and will make every provider/hoster stronger and benefit consumers. I think it will increase competition rather than hinder it. If a provider chooses to be 'in' they better be in past their earballs. If they choose not to be in, they need to clearly articulate the value of why they chose their path. The market will not tolerate lip service to being 'in' nor MABUSHI reasons for not being 'in'.

James, consider that 1) there

James, consider that 1) there are few providers running openstack yet. That ecosystem is likely a year out given the time service providers take to stand up a new service and 2) once they do stand up an openstack based service they will add a significant amount of proprietary code to integrate with their existing OSS.

The real advantage openstack brings is the cost model for providers is now compelling. But I would not expect to see an ecosystem of cloud providers working together quite as readily as might be expected.

Ray Nugent

re: OpenStack ecosystem a work in progress

A fair comment, Ray, but note that there are already providers standing up IaaS based on OpenStack components - NASA Nebula, Rackspace, NTT, KT and others and we already have evidence of their efforts to collaborate.

Certainly there are a few

Certainly there are a few early adopters. Remember thought that NASA is not a provider, RAX has deployed very little openstack functionality yet and the other two are outside the US and only partially operational. Openstack has a lot of promise, my point is that CSPs may not cooperate to exploit all of that promise.

Ray

Operations iterations are also different

It's not just about code visibility, OpenStack will be deployed in many different ways, so we will find right ops models faster. AWS has to go it alone.

RobHirschfeld.com, @Zehicle

Promise of The Cloud

It's an interesting comparison between the two and maybe the open source path may help some reluctant IT folks move forward with the Cloud but....the Cloud's claim to fame is the abstraction of the complexity and management of said complexity away from the end user. Certainly taking a look at the source code or heaven forbid modifying it is not exactly getting away from the complexity and is in fact beyond the capability of the average IT shop(not all just most). I know a lot of this discussion is coming up because of the AWS outage and that is probably valid. It seems though that Amazon provided enough information for some Architects to have had their solutions survive without issue while others did not. I not sure that having the source code would have helped the ones that did not.

markg
http://www.thegreylines.net