We recently had the pleasure of having an extended conversation with Antonio Piraino, CTO at ScienceLogic. Our chat covered a wide range of topics, including the cloud industry at large, enterprise adoption of cloud, and the differences between the developments in cloud technology and the developments among cloud service providers.
This is the first of 2 parts, so check Part 2 here. Let me know your thoughts on Twitter @CloudGathering.
Gathering Clouds: What are your perspectives around how the cloud industry broadly has developed over the last few years?
Antonio Piraino: Having been a former analyst, there’s so much to say on where it started because I was already an analyst when Amazon Web Services (AWS) was getting going and started in the United States with Simple Storage Service (S3) and then Elastic Compute Cloud (EC2). It started getting really interesting when they started introducing all of their new services like Elastic Block Store (EBS) and the Simple Queue Service (SQS) messaging services and a whole bunch of their other services as well. It’s funny because it took a while; they were almost untouchable for a couple of years and then you could almost see Microsoft waking up when they launched what’s now Windows Azure. You could see Google get it as well and then launch App Engine, because even though Google App Stores was just coming out at the same time into mainstream, I think the company realized that there was a concern; they didn’t want to miss out on the opportunity to go after the enterprise sector which was still more focused on the infrastructure layer and less focused on buying a software-as-a-service (SaaS) type pre-packaged application offerings. So I’ll give credit to Amazon and a lot of the other big guys for leading the way when it came to modern day cloud computing. Even though hosting had already existed for a very long time, fundamentally what changed first and foremost was the ability to get billed by the hour instead of having to take out a two year contract and pay by the month.
A lot of people forget that. But I remember the first time I spoke to Amazon and I said, “What are you doing differently than the average hosting company out there who’s operating a splice of a server or a shared hosting environment or a managed hosting environment?” And they said, “Well, first of all, it’s a billing model change.” And they were absolutely right. And that took the world by storm because that’s what made so many developers so eager to use them. Amazon was really driven by the development community and the realization that they didn’t need as much management; that they didn’t need a framework for development so much as they just needed the tools and the resources to run their code on. That’s really where it took off.
I give kudos to the managed hosting providers. They figured out very quickly that they needed to get there, too and start becoming a lot more dynamic, not just in the availability of the resources on a more granular basis. Some of them started that way, Rackspace and a couple of other jumped on the ball and said, “Can we make available resources and billing on an hourly basis, too?” But I think they very quickly realized that the more mature enterprises weren’t necessarily in love with that hourly payment schema. But it was many of the enterprises that had led the push towards virtualization. And, more importantly, after a couple of years of operating on virtualized environments these hosting providers also realized that the bigger opportunity was in the tools that came with virtualization, the flexibility; the ability to do snapshots, cloning and roll-backs; and all of these second order events that were far more convenient for the average IP administrator than they’d ever had at their fingertips in the prior decades. So, suddenly, there were these new tool sets that cloud computing stimulated, which are, I think, almost more important than the fact that cloud computing even became a recognized paradigm in the first place.
GC: What’s your perspective on how larger companies like Google, Microsoft, and HP have entered the cloud market at a scale that puts them in competition with Amazon versus how the managed service providers have focused their differentiation to Amazon?
AP: There’s a lot of convergence, and I’ll use that word very broadly. There’s a convergence in thinking; a convergence on… the target market that everyone’s going after; and convergence in the IT infrastructure itself as well. There’s even convergence, frankly, in the way in which IT operations happens; and convergence in the data center. There’s just so much convergence going on across the board. It’s coming together like a perfect storm, as a lot of people have talked about, and the realization is suddenly dawning on people because the big enterprise sector was really the laggard in the space. They took a long time coming to cloud computing. When you hear the Gartner’s of this world talking about “What is cloud computing?” and “What are the enterprises’ view of SaaS?” – that’s what we were talking about five, six, seven years ago amongst small businesses, the IT world, and hosting providers. So, they’ve really been slow to jump on the bandwagon.
But enterprises represent a much bigger opportunity. Facebook, Google and Amazon have done well in the consumer sector and in the developer community, and I can’t wait until we see what they do in the enterprise sector because there’s no way they’re going to miss out on that opportunity. Microsoft, HP and Google are building data centers today and those data centers are not simply being built for fun, or to continue supporting what those companies are doing today; they are really going to go after serious business applications and opportunities. Managed hosting providers – those who are on the ball and at the forefront – have realized that opportunity, too. Their differentiator is not going to come by trying to undercut the mega clouds in price; they’re going to have to do it through high-touch management of their customers, and also through a higher level of services. They’re not going to be able to purely compete at an infrastructure layer anymore because infrastructure-as-a-service (IaaS) has become very commoditized, as are platform-as-a-service (PaaS) offerings for developers. So, they have to start offering more unique services, more unique applications, more unique management and control for their customers; these are the big differentiation points that are permeating right now in the cloud world.
Providers no longer say, “Hey! We’ve got an infrastructure that’s available to a not very well-educated customer who’s not going to do the homework beyond just one or two providers, and just like the deal better than their internal IT environment.” Now it’s a much more educated customer who’s saying, “Hold on; I have an option here because VMware doesn’t just sell me software; they sell me a public cloud, too.” HP provides the same thing. So as a consumer, you have a lot of variable options at your fingertips that you are now aware of.
The consumer’s challenge is that they have to be much smarter about the way they do IT. So, they’re looking for much more granularity than they did before, at least in the United States (US). I would say the rest of the world is a little bit behind from where the US was a few years ago, and most non-American consumers are probably currently more driven by price reductions and convenience than anything else. Whereas now in the US, we’re in the world of “How much more convenient can this be? How much more efficient can this be for me, cost-wise, but also productivity-wise?”
GC: What are the biggest challenges facing cloud as a concept or cloud as a paradigm currently?
AP: Very interesting question. It’s very, very easy to take a US-centric approach, and there are significant differences between the US and Europe. One of the key differences is that Europe is 98% small businesses. Even mid-sized businesses in Europe would be considered small business in the US. And I think that does lend itself to a different set of concerns; a different set of criteria for acquiring cloud computing altogether. But putting back on my US‑centric hat for a second, the big challenge that’s facing the big service providers is the fact that the majority have not done enough to elevate themselves above the fray and it’s very quickly becoming commoditized. So you can call it hosting; you can call it shared hosting; you can call it managed hosting; you can even call it cloud computing. And even though I believe in the straight definitional differences between them all, I’m not sure that the rest of the world particularly cares about what the difference is; they are simply trying to solve a problem in their business, be it capital expenditure (CAPEX) or on IT infrastructure overall.
I think more and more so, cloud is resolving the operational cost element. Particularly in the US, among small businesses, it’s about doing things faster, more efficiently and more conveniently because if you are a small business, you have fewer, smaller, and less sophisticated IT departments. These employees cannot possibly be experts in security and in hardware maintenance and virtualization and in every application that goes on top of the virtualization layer. It’s just too much to take.
The opportunity for cloud service providers is to show consumers that they have those economies of scales, as well as having all the different tools available. The challenge at the same time remains very much associated with IT managers. More often than not, the head of an IT organization, particularly a large IT organization, will want to build solutions themselves and to be viewed as the expert for all things IT and computer-related.
And that’s a big challenge to get over one’s ego as a cloud provider and say, “Listen guys, as you grow, from small to big organization, your solution is just not going to scale because you just have not put the man hours and focus on building a software code that’ll answer all the questions that you want to get to. If you don’t, then your competition is now deploying.” And this gets down to your absolute advantage. It’s just like macroeconomics. Most countries should really be building a product or a service. It is not in their absolute advantage to build something when another country can do it better; they should focus on what they do really, really well and then trade that service or product for what the other country does very well. Similarly speaking, any enterprise should really be focused on the business that they do really, really well. And it doesn’t mean they have to fire their IT staff. It just means that the IT staff can be used more productively on things that matter. And convincing these companies of that fact remains the biggest challenge for cloud service providers; that enterprises and small businesses don’t have to do it all themselves. I think small businesses continue to be more financially sensitive and big business are probably far more focused on robustness, service-level agreements (SLA), speed of deployment, and fortunately, the kind of technology used, since there is still an association with using big brand names.
GC: Over the last few years, what are you seeing with regards to growth for outsourced cloud service providers, and where do see this growth headed? How are service providers differentiating themselves and how are they keeping up with the rapidity of change in the industry overall? Where are they going to go next with what they are providing to clients as a value-add?
AP: Everybody talks about how enterprises are emulating service providers in their internal practices. But I actually think that service providers have something to learn from enterprises, too. Enterprises have been spending a lot of money on virtualization, just like the service providers have done. But where I see the enterprises now going is trying to spend a lot more money on automation and orchestration tools to manage those virtualized environments, because they realize that they’re starting to lose control of it all.
If I were a big business today, my expectation for my cloud provider wouldn’t just be to take away any CAPEX, but also to remove operational burden, and to have a far more cutting edge bit of technology than I have in-house. And that means a service provider has to be as robust as my business is, if not better. On top of this, the service provider should be as secure as I am, if not better because I don’t want to have to worry about that. All of this should be far more automated with the provider as well.
That’s where I think the real opportunity lies: being more automated so that if you have policies that ensure that not only do I not have downtime, but that I also have the right resources being used for my workloads and I’m only being charged for the actual utilization of those resources. Cloud providers are really getting a lot more efficient both on the billing side of the resources but also on the automated provisioning. I think that’s where service providers are headed next: toward a lot more investment in automation, orchestration and in management tools. At the end of the day, companies like Rackspace are going have a very hard time. As a provider, you want to be high-touch, in the consultation as well as in the upfront sales processes.
But cloud providers have to start creating far more automation, and that’s where the next big wave of really strong cloud computing providers is going to come from: folks that are doing all these things that the managed services providers are doing but have automated all those processes so they can start focusing more on the delivery of high-level applications, services and bonding together different applications that depend on each other.
GC: What does the future hold for managed service providers?
AP: I don’t think that there’s much more room for entry into the cloud computing provider market. And when I say cloud computing, I include all data center service providers and hosted and managed service providers. I really don’t think there’s much room for new entrants to really come in and “wow” the market. I think if there was going to be anybody who wowed the market, they would come from an existing large organization, but I don’t think there’s entry from the bottom as a brand new start-up very easily into this market.
I’ve seen firsthand where thousands upon thousands of small service providers have become lifestyle businesses: they grow quickly at the beginning because they get a handful of anchor tenants who generate their revenue. But to really grow large means spreading geographically beyond your immediate customer base, because most businesses tend to have local customers. Who we see in the market today among the cloud providers are going to be the leaders tomorrow; notwithstanding the fact that Amazon, Google and Microsoft will continue to “eat our lunch,” so to speak. And quite frankly, that is still a possibility and a real danger. The danger will manifest itself between now and the next sort of 8 to 10 years, but I think it’s a real danger.
Having said that, I think that the success of these service providers is really – I know it’s not a great answer – going to be measured on a dollar-by-dollar basis. In the end, success is in the margins.
GC: For the enterprise looking at cloud computing, what are the chief concerns that still exist? What are the hurdles that you see for the enterprise?
AP: Security is a concern, but it’s not a hindrance for adoption. I think every service provider has to have an answer for it and I think enterprises are willing to accept those answers; they just want to know that there’s been some thought around it. The onus is on the enterprises to know what they’re spending their money on. They need to be aware of the fact that there are service providers and there are service providers; just as there are cloud providers and then there are cloud providers. If you’re willing to go and buy an instance at AWS and not bother to look at the SLA or buy some redundancy in multiple zones, then you’re asking for trouble.
Enterprises are not going to be held back because of security concerns because the right security people do exist out there. There are the Carpathias and the Terramarks of the world. They, and Logicworks as well, have an appropriate level of security that they will keep updated beyond what the average mid-sized business will be doing themselves.
The next big wave of opportunity is now arriving with the enterprises; that cloud computing is very palatable to them and that they simply need to check the box where security is concerned. What is rapidly changing is that outsourcing of IT is no longer just being led by the underlining infrastructure, but much more by the applications and services such as video conferencing, for example. Other collaborative and business productivity applications like hosted desktop are really the next pathway into these enterprises.
GC: Where do you see innovation happening?
AP: As an innovative software company ourselves, I would have to answer that question in a very biased way; but… but let me try to be somewhat objective about it.
There’s perceived innovation and there’s real innovation. What I mean by that is that I’ve talked about convergences in the stacks. You see NetApp working with VMware, VMware working with EMC, and so forth. Everywhere you look, you see these converged infrastructures.
And truth be told, I haven’t seen any compelling evidence from anybody who has deployed them, that these are truly innovative, crazy “changing my life” type of innovations. But at the same time, this stuff is selling like crazy and being deployed all over the place. So, that’s what I mean by what’s perceived and what’s real where innovation is concerned. I truly think that the real innovation is actually happening much more amongst software providers, since that’s where the innovation is much faster.
Virtualization has become commoditized, the hypervisor is commoditized, and even amongst the hypervisor vendors, if you look at Cisco, Microsoft or VMware for example, much of their concentration on innovation is happening around the orchestration layers and the manageability layers because they’re realizing that they have to start managing more. VMware recently announced that vCenter is now managing up to the application stack. There is the recognition that they have to start managing increasingly up and down the stack, and even beyond it in many cases. These companies can’t help but manage their own technologies, of course, but they have to start being broader about how it’s done. In instances when the current market leaders can’t quite get there, there are other innovators who are helping to fill the gap, in manageability, orchestration capabilities, provisioning capabilities and so on.
Innovation is becoming more possible as well because the entire world seems to be opening up a little bit more in terms of the architectures and the API’s that are becoming available. API’s are more open and accessible in allowing more innovation from third parties, too. vCloud Director is just an example of that. In speaking of inventors of multiple APIs, ScienceLogic, for example, is able to come in now and suddenly play more of a role in orchestration than we ever had before when we were purely viewed as a monitoring and eventing type of company. Innovation across the board is being driven through partnerships. And I think there’s a greater realization that we all float or sink together, so there’s no point in closing things off in a completely proprietary manner.
By Jake Gardner