Every time we talk about mobile devices the discussion goes directly to Apple and its ecosystem as the key differentiator for success. Can we say the same thing about cloud?
At the moment I’m carrying out a small project and I’m doing some tesst with different providers and the deeper I go the more I understand that when you use these kind of tools and services the difference is in the “user experience” that comes from the ecosystem.
Clouds and mushrooms
You know, new CSPs are popping up like mushrooms promising new cloud experiences with brand spanking new services, but every time I try them out I’m disappointed.
Often, just obtaining a pre baked (certified and optimized) WordPress instance is quite a job in itself (I’m not the kind of person that usually spend hours to install, compile, configure such a LAMP stack… I take it for granted that someone better than me in sysadmin tasks base already done it somewhere!)
They justify themselves with the fact that the services are just in the early stages while mentioning that soon they’ll be adding features, a PaaS offering on their IaaS, partnerships with other vendors to offer Virtual Appliances, and blah, blah, blah. But I can say that this process is slow and usually, not developed at their best…
From the end user standpoint the issues are always the same: “Openess” or “hadcraftable” are not very important characteristic if compared with overall usability and satisfactory user experience.
Amazon has an ecosystem, which is probably the best one that you can find nowadays on the public cloud global offering. It’s not “open” but it’s easy to use, feature rich, coherent and it also has a store (AWS marketplace) with a vast choice of AMIs (AMIs are the Amazon’s VM templates) that you can use for free or pay with the same pay-as-you-go model used for all other services!
Actually, Amazon too has its own flaws (even when some architectural design aspects could be considered features or defects). But AWS is a place where developers and third parties could easily monetize their efforts while end users can have a coherent and solid experience.
One aspect that I particularly love very much is third party enhancements tools (I could coarsely defined them add-ons or plug-ins) that are often developed on AWS to improve its functionalities (i.e.: here and here).
Vmware isn’t Amazon and vCloud isn’t an AWS competitor but Vmware is building a cloud ecosystem around vCloud. This ecosystem is different but interesting to me.
Vmware is working a lot to push some projects (eg. here and here) aimed to create a useful near-out-of-the-box infrastructure ecosystem for who wants to deploy a private or public cloud service (they are focused on the builder rather than on the user). Wmware is a software vendor and theoretically (and practically?!) it isn’t in direct competition with Amazon AWS but the competition is between AWS and VMware customers.
From my point of view it’s very interesting to look at VMware: they interpret the cloud in a more “traditional” way and they are very focused in shifting legacy infrastructures/apps to the cloud (smoother than the competition).
My readers know that I love Joyent approach and, actually, they are going in the same direction of the firsts two, building something in the middle between AWS and vCloud: they have a public cloud ecosystem (limited if compared to AWS) and they also sell their infrastructure software to build clouds. Joyent has an ambitious roadmap and I’m keeping an eye on them for different reasons.
And that’s not all, many other international players like Rackspace, only to name one, are working in this direction (also if project OpenStack still leaves me doubtful). And I must admit that I hear many names every day that are working on similar projects but never in the right manner or context…
Invest in ecosystems
Amazon has invested a lot in its cloud but not only in the infrastructure (and marketing) they also spent to build a coherent ecosystem and on the evangelization in using it.
The provider (also the local one) must invest to build its ecosystem if it wants to compete… Probably, from this point of view, Openstack-like solutions (it they reach a certain level of maturity) could become the right choice for small CSPs. In this case the investment in the infrastructure is limited while the CSP could spend its money to:
1) educate users and developers about features, APIs and the “open” AWS-like vision
2) involve partners and developers to build a marketplace where both can earn money on optimized appliances.
In summary, the end user wants an experience in line with his expectations… People don’t like to go back: they expect to have a totally new experience (like Amazon’s AWS), where they are willing to change to their habits, or they want a rock solid experience (like Vmware vCloud) where they can obtain the same experience that they remember from their ordinary virtualized infrastructure.
If you show a limited cloud with the ability to only spin up a VM and perform basic operations on it, (like 10 years ago with first hypervisors) obviously that user experience will fail.