Overcoming cloud complexities
Cloud infrastructures 'an asteroid belt' to navigate - chat with Zenoss's Floyd Strimling
Firstly, do you feel it's an exciting time for cloud infrastructures?
It's an exciting time to be in IT in general as we are seeing breakthroughs across the board. From solid state drives to falling price of memory to powerful compute platforms, to new network paradigms, it's simply an amazing time.
Within "Cloud" specifically, the pace of innovation and competition is driving a revolution that is breaking down traditional silos while offering unprecedented agility that enables both traditional and nontraditional business models to thrive.
Do you think there's simply too many options for enterprises nowadays?
I hate to say this, but there are almost too many choices for enterprises. enterprises are dealing with critical infrastructure that has been deemed to be "legacy" by the likes of Facebook, Google, and more. To keep these systems up and running while exploring the multitude of options for a next generation IT infrastructure is a monumental task. Not to mention the fact that nearly every vendor is marketing a similar story.
Is it a minefield for enterprises to get exactly what they want in terms of security, frameworks and the like?
I wouldn't call it a minefield, more like an asteroid belt. Enterprises must navigate a maze of challenges that they have not yet overcome within the traditional IT model while moving to the cloud era. The need for security, frameworks, and more is consistent across both paradigms albeit with more fear around the cloud. The real issue lies within the disruptive nature of cloud itself that threatens the very vendors that these enterprises have come to rely on for consultation and solutions.
What are the main issues for a business when choosing a cloud infrastructure? What should they consider?
In reality, every business will need to ask themselves these questions prior to moving to the cloud. Last year I advocated for an Enterprise Cloud Triage (ECT) to help enterprises determine their readiness for cloud. This includes understanding the business objectives, people, an assessment of your current infrastructure, and exploring various doomsday scenarios. In any case, the number one thing an enterprise should NOT do is rush to the cloud or make arbitrary technology decisions that will impact them for years to come.
To truly become successful, does a business need to manage cloud complexity effectively quickly?
To me, the number one issue facing cloud computing is complexity. While security or continuity will be solved, the fact is the cloud paradigm isn't as "cloudy" as people would like. While the complexity has been reduced for the consumers of the technology, but from the administrators perspective things are getting more complex.
Learning new technologies like OpenStack, CloudStack, Puppet Labs, Chef, and more may be easy for the IT elite, but it's not quite that simple when you are managing 1000s of critical applications within today's always connected world. For enterprises to be successful, they must include a viable strategy for their IT Operations to manage both legacy and cloud complexity.
It's difficult to use a tool before it becomes already out of date and sufficient. How does an enterprise combat this problem?
This is exactly why the open source revolution is so powerful. Why purchase a proprietary solution that requires an army of consultants to implement and maintain it when you can utilize modern and agile commercial open source solutions? The key is the solution must be open enough to allow enterprises to take-on today's challenges with an eye to the future while being supportable and upgradeable. The downside, is enterprises will need to step out of their comfort zone to implement solutions with new names like Zenoss, Opscode/Chef, Puppet Labs, Piston Cloud, and more.
What did you make of Citrix's decision to jump the OpenStack ship and head to Apache with CloudStack? On the cards?
To answer the question, I'd have to believe that Citrix has abandoned OpenStack which I don't believe is the case. Citrix continues to contribute to OpenStack and they are the custodians of Xen. With that said, CloudStack deserves the opportunity to become a top level open source project under the Apache Software Foundation (ASF). Whatever happens in the future, I believe it's a brilliant move for Citrix and gives them a real chance to disrupt the cloud market. However, simply moving to ASF doesn't guarantee success and Citrix has a lot of work to do to make the CloudStack project successful.
Similarly, what do you make of Red Hat's big backing of OpenStack?On March 1, 2012, I wrote a blog post entitled, "Why Billion Dollar Red Hat and OpenStack Need to Dance", which broke all of Zenoss' blogging records. The fact is, I understand why Red Hat originally rebuffed OpenStack but I'm thrilled that they were able to reconcile under the OpenStack Foundation. Red Hat could not afford to lose the cloud to other rival Linux operating system vendors such as Canonical and they have an opportunity to create a commercial version for the enterprise that is fully supported and tested.
Will Red Hat at the table truly push OpenStack forward now it has that huge partner? And what does it mean for their other products ie OpenShift/Cloudforms?
Now that the inevitability of Red Hat joining OpenStack is over, it's time for Red Hat to get to work and deliver a commercial version of OpenStack to the Enterprise. It's my opinion that this move will force a strategy shift within CloudForms as I doubt Red Hat will want to maintain multiple IaaS platforms. However, OpenShift may be different a la VMware's Cloud Foundry whereby a PaaS solution doesn't necessarily need to conform to a single design paradigm.
Is there space for both CloudStack and OpenStack in an already stocked marketplace? Can they co-exist? Who will be the winner here?
To me the question isn't really about CloudStack and OpenStack but their commercial offerings by 3rd party software and hardware vendors. Will OpenStack or CloudStack be deployed in their open source forms within an enterprise without the backing of a commercial entity? Perhaps within the technical elite, but the majority of the Enterprise they will look to a commercial offering.
Therefore, the real question is who will challenge Citrix's commercial offering with a viable commercial OpenStack solution. Today you have Piston Cloud, Nebula, Cloudscaling, and of course Rackspace with the eventuality of a Red Hat offering. These solutions will need to effectively compete against the likes of VMware vCloud and, even more interesting, Eucalyptus with their Amazon Web Services (AWS) relationship.
Do you feel not enough enterprises are embracing different models - public, private and hybrid cloud-based infrastructures?
Enterprises have always been open to new models within IT infrastructure. Whether it was moving from the mainframe to client/server or now to the cloud the business drivers must trump the technology. As I talk to large and small enterprises, they are all investigating cloud-based infrastructure but they are all struggling with the journey itself.
However, the enterprises that embrace new technology and paradigms will have the ability to compete against their current rivals as well as unseen challengers. Each enterprise will need to decide which model fits their business goals and objectives and I'd suggest they start small and grow into the cloud. After all, that's the benefit of a scale-out and agile infrastructure design.