A couple weeks ago, I swore never to use 'multi-cloud' in a blog post, and here I am writing about it. I hold CliQr, a multi-cloud inspired startup with a new approach to cloud provisioning responsible.
For a young startup, CliQr has diverse use cases I am still bending my mind around.
Before I get to what CliQr is up to, there is a more fundamental question: what is multi-cloud, and why should we pay any attention to it? Also pressing - what customer problem does CliQr intend to solve?
The first time I saw a reference to multi-cloud was in a blog by Ben Kepes, The Multi Cloud Opportunity – Delivering on the Composable Enterprise. Reflecting on a panel from GigaOM's Structure event in June, Kepes grappled with why we need to add multi-cloud to the growing list of cloud buzzwords.
Kepes floated a distinction between 'hybrid clouds,' which are often used to describe a mixture of one private cloud and one public cloud resource, and 'multi-cloud', which seems to imply not only multiple cloud environments, but more flexibility. Translation - less cloud vendor lock-in:
Multi-cloud on the other hand is a far more mature and pragmatic approach towards cloud infrastructure and, in my view, directly delivers upon the “composable enterprise” notion that Warner Music group CTO Jonathan Murray spoke of at the event. An organization embracing a multi-cloud strategy will likely have infrastructure resources spanning multiple vendors, potentially public and private, and likely to be across a variety of operating stacks. In the multi-cloud scenario it’s not a shock to see organizations using AWS, some public OpenStack and a private cloud build on one of the stack products.
My first response: sounds empowering to customers. But it also sounds like a headache to manage, no?
I asked Esteban Kolsky, a fellow Enterprise Irregular who also attended the CliQr briefing, about the cloud issues enterprises are facing. He responded:
Two trends organizations are noticing as they move to a world of public cloud: the need for multiple providers to ensure scalability, reliability, and competitive costing, and the need to build their own applications internally leveraging these resources. There are no easy answers to this, it is too early for most large and established vendors to even begin to realize these needs exist. As always, those solutions come from the fringes, from startups committed to doing one thing very well as opposed to try to control too many. CliQr is one of those vendors.
Later in his multi-cloud post, Kepes cited Paypal and Box as two organizations that are making an important discovery: different clouds have different advantages for different workloads.
Cloud realities - CliQr's view
CliQr provided me with a few slides that frame the opportunity. The first point? The enterprise cloud is real - and growing fast:
Given the recent earnings reports of cloud companies, I don't think anyone would quarrel with this premise (I have recently blogged on cloud adoption in the large enterprise). The next point I would describe as the 'challenge/opportunity': cloud diversity and use cases are increasing. Headaches on the one hand, opportunities on the other:
So what does CliQr intend to do about it? First, a few factoids: CliQr is a post-series A company with cited investors Google Ventures and Foundation Capital. Their development happens in India and the U.S. via one month sprints. Version 2 of their enterprise-ready platform has been available since April. In a nutshell, besides providing enterprises management and control across private/public clouds, CliQr aims to provide true portability of applications across clouds.
CliQr has almost a dozen customers in the field now, and while I can't cite any customer names here, I will likely be interviewing one of their customers in the future. Potential customers include both enterprises and ISVs. The use cases vary from practical to visionary.
A practical example: an ISV could use CliQr to provide 'cloud portability' services to their end customers, transitioning their customers onto a cloud of their choice or moving them to a different cloud provider (CliQr's menu of cloud providers includes most if not all of the usual suspects, you can see the rundown on their home page).
The sizzle - application-specific cloud benchmarking
There was one CliQr use case that stood out: application-specific cloud provisioning and benchmarking. Or, in plain English: What's the best cloud for the app?
Bottom line? Cloud performance varies greatly on a per-app basis. What if you could benchmark application performance on different cloud providers and run apps on different cloud providers as needed? That's a core part of the CliQr model.
As CliQr CMO David Cope put it during our briefing: 'We have the ability to benchmark an application across different clouds, I can for example look at existing benchmarks that we did on the same application across HP, Amazon, and Rackspace.'
The specifics of the benchmarking are where the real meat is. Cope:
How long does it really take to complete a job? You can measure that level of benchmark. Or you can measure the computer request time per second. Now you ask, what can I do with this information? Well, instead of just choosing a cloud, you can say decide which criteria do you want to work with on an application-specific basis. Do you want to choose best price, do you want to choose best performance, or best price for performance?
The CliQr team acknowledged that not every customer is ready for application benchmarking. But in terms of early adopters, they cited oil and gas and telecommunications as two industries where high performance computing needs can add financial urgency to this kind of benchmarking.
High performance computing scenarios (like gene sequencing) entails large batch jobs where the ability to choose the right cloud and run a benchmark has immediate impact. A customer example cited on the call involved a manufacturer that needs to be able to run product simulations in a live environment. But running hundreds of nodes adds up to a big invoice pretty quickly. As Pradhan put it, 'Before they run their simulation, they are fully capable of making a decision on, 'Should I run this in HP Cloud, or should I run it in Amazon? If I run it in Amazon, which region should I run it on?'
Views on CliQr's potential
Esteban Kolsky is not easily impressed, so I was honestly a bit surprised by his optimism on CliQr. After the briefing he emailed me this comment on CliQr's prospects:
CliQR has a very interesting approach to analytics that can give IT departments perspectives about their use of cloud resources (and the most efficient way to manage them) as they never had before. An early-stage startup that is still figuring out their message and best go-to-market strategy, I am impressed with what they have done and their understanding of the value an open cloud provides that I will keep an eye on them to see their progress. In spite of being early to market, their timing is superb as most IT organizations are just now beginning to realize that managing myriad private clouds is not a solution they can sustain over the long run, and they begin to understand the demands their stakeholders are planning for their cloud infrastructure.
Louis Columbus, another Enterprise Irregular on the briefing, told me afterwards that he sees potential in use cases involving regions where cloud data security is an issue:
I have been thinking over that briefing from the perspective of how their CloudCenter Manager working in conjunction with the CloudCenter Orchestrator have the potential to unlock new markets for cloud computing where compliance (i.e. HIPAA) and data ownership requirements imposed at the national level (like Germany) are barriers to entry today. CliQr's potential in those scenarios are interesting to think about.
For what it's worth, CliQr confirmed on the call that their offering has been well received by German companies so far.
I still don't care much for the multi-cloud buzzword, though I can see why such terminology might be needed to underscore more customer choice and flexibility than the 'hybrid' term implies. That second part, about customer choice and provisioning flexibility, I am much more bullish on - though clearly not all customers are ready to take on such chores, not matter how user-friendly the tools. Not to mention that countless terrific products have died on the enterprise vine because of go-to-market challenges. The sales and marketing side of this is another discussion for another time.
Whether CliQr has timed the market right remains to be seen, but the notion of application-specific cloud benchmarking does seem promising. I'm not sure if this gets us closer to the 'cloudbursting' scenario Kepes has dismissed as marketing 'unicorns and rainbows' (cloudbursting is the concept of frictionless shifting of resources between public and private clouds).
But it's an interesting conversation to have, and with startups like CliQr pushing the cloud performance envelope, we may be spotting unicorns sooner than some of us thought.
Image credit: Businessman standing in a desert © WavebreakmediaMicro - Fotolia.com
Disclosure: CliQr is not a client of mine nor does CliQr have an affiliation with diginomica. The Enterprise Irregulars is a loose consortium of enterprise influencers and practitioners. Phil and I are EIs; Den was a co-founder.