The next edition of the HP Discover Performance Podcast Series brings together three HP executives to explore the implications and business value from the Converged Cloud portfolio updates announced at HP Discover 2013 in Las Vegas.
There's no hotter topic — and nothing more top of the mind these days — than cloud computing. Not surprisingly, HP has made that a major focus at day two of Discover.
To put some context around it all, BriefingsDirect assembled chief evangelist at HP Software, Paul Muller; Christian Verstraete, chief technologist for Cloud Solutions at HP; and Tom Norton, vice president for Big Data Technology Services at HP. The panel was moderated by Dana Gardner, principal analyst at Interarbor Solutions.
Below are some excerpts.
Christian, tell us about the state of the market before we get into HP's response to it.
Verstraete: What's happening in the market today, is that on one end, you have startups that are rushing to the cloud very quickly, that use cloud and don't use anything else, because they don't want to spend a penny on building up an IT department.
On the other extreme, you have very large corporations that look at all the things that are unknown around cloud and are sticking their toe in the water.
And you have everything possible and every possible scenario in the middle. That's where things are getting interesting. You have forward-looking CIOs who are embracing clouds, and understand how cloud can help them add value to the business and, as such, are an important part of the business.
You have other CIOs who are very reluctant and that prefer to stay managing the traditional boat, if I can put it that way, in keeping and providing that support to our customers. It's a interesting market right now.
Muller: The challenge that both vendors and consumers have is that one size does not fit all. When you find yourself in that situation, you only have two responses available to you.
If you're a one-trick pony, if you have only got one technology, one approach, then it's one size fits all. Henry Ford, one of your fellow countrymen, once said that you can have the car in any color you want, so long as it's black.
It's a great idea in terms of simplifying what your choices are, but it doesn't help you if you're an enterprise that's struggling to deal with complexity and heterogeneity.
We believe that there are three absolutely critical priorities that anyone looking into cloud should have. The first is confidence. Confidence, because you are moving typically mission-critical services in it. Even if it's develop and test, you're counting on this to work.
The second is consistency. There is absolutely nothing to be gained by having a cheap cloud service, on one hand, and then having to retrain people in order to use that, because it's completely different from your internal systems. It's just moving costs around. So consistency is absolutely critical.
The third piece we talk about all the time, choice. You should have your choice of operating system, database, and application development environment, whether it's Java or .NET, you shouldn't have to compromise when you're looking at cloud technology. So it's those three things — confidence, consistency, and choice.
Giving users choice
Here at Discover, we're hearing a lot about a variety of announcements for Converged Cloud. Let's look at a couple of these major aspects of the announcements and then delve into how they come together, perhaps forming a whole greater than the sum of the parts.
The first part, Christian, is this real emphasis on OpenStack and the Cloud OS. So give us a quick overview of where HP is going with OpenStack and Cloud OS, and how that relates to some of the requirements that we've just discussed?
Verstraete: What we want to do across the different clouds that we offer — private cloud, the managed cloud, and the public cloud — is a capability to be able to port workloads very quickly to build some consistency around them.
Cloud OS is all about that. It's about building a consistent infrastructure environment or infrastructure management environment to do that. And that's where we are using OpenStack.
So what is cloud OS? Cloud OS is nothing more than HP's internal OpenStack distribution, with a set of additional functionalities on top of it, to provide a second-to-none infrastructure-as-a-service (IaaS) delivery that can then be used for our private cloud, our managed cloud, and is already used for our public cloud.
That's the first thing that we announced. We are building on top of that. It's an evolution of what we started about a year and a half ago with Converged Cloud. So we just keep moving and working around with that.
We also announced that we not only support Cloud OS in our traditional blade environment and our x86 servers, but also on the newly announced HP Moonshot servers. That combination may become interesting when we start talking about the "internet of things" and a number of other things in that particular area. It will also provide our customers with the capability to test and play with Cloud OS through a sandbox. So there's a lot of emphasis on that.
It also seems that you are expanding your support of different virtual machines (VMs), so heterogeneity is supported. As Paul Muller pointed out, it's supporting all the various frameworks. Is there something fundamentally different about the way HP is going about this cloud support with that emphasis on openness vis-à-vis some of the other approaches?
Verstraete: Many of the other players, many of our competitors, have what Paul mentioned earlier, a one-trick pony. They're either in the public space or the private space, but with one hypervisor. Where we're starting from, and that's the essence of Converged Cloud, is to say that a company going to cloud is not one size fits all. They're going to need a combination of different types of clouds to provide, on one hand, the agility that they need, and, on the other hand, the price point that they're looking for.
They'll put some stuff in their private cloud, and they'll put some other stuff in the public cloud. They'll probably consume software-as-a-service (SaaS) services from others. They'll probably put some things into a managed cloud. It's going to be a combination of those, and they're going to have to handle and live with that combination.
The question is how to make that easy, and how to allow them to access all of that through one pane of glass, because you don't want to give that complexity to the end users. That's exactly where Cloud OS is starting to play. Cloud OS is the foundation for us to do that.
Tom Norton, seeing that this field is very diverse in terms of the needs and requirements, it seems like a perfect fit for lots of consulting, professional, and support services, but we don't often hear about them in conjunction with cloud. Tell us a little bit about why the market is ripe for much more emphasis on the services portion here?
Norton: As you start to take advantage of the varying services that are available through the cloud, or that you want to present to the cloud, the varying presentation formats, the varying architectures are an issue for whether you're a startup or in the enterprise.
From a consulting perspective, you need to have a strategy and understand the challenges and complexities of that hybrid type of delivery or that hybrid consumption, and establish some type of design for how that's going to be used and presented. So consulting becomes very important the more you start to consume or present cloud-based-type services.
When you start thinking of that design and that whole approach from balancing across the network to balancing the infrastructure component pieces, you need to have some kind of consistent support structure. One of the most expensive parts of this is going to be how you support those different environments, so that if you have an issue, you're not doing component-based support anymore. You need a holistic-based cloud support.
For organizations truly to transform themselves as an IT organization and be able to present their service, which in many cases is an application, that app may be something presented internally to business units because the business units are getting some value, or even externally to a customer or to a customer's application.
Those apps are designed, in many cases, in either a more mainframe-based environment or also in the distributed environment. When you start thinking of presenting it as a service, there are other considerations that need to take effect.
You start looking at how that application performs in terms of more virtualized and automated environments. You also think about how you can manage that application from a service perspective. How do you monitor the application? How is it metered in terms of the presentation? How is that application presented within a service portfolio or a service catalog? How do you then manage and monitor the application for service operations? The user demands an end user experience for meeting a certain service level.
When you think of modernizing applications to a cloud-based presentation, there are multiple layers that have to be considered to even address the applications. When you think about the application piece and the work that needs to be done, you also have to think about the management component pieces of it.
That's why you'll hear of services around, say, cloud design services that will enable us to take a look at that service portfolio, look at the service catalog, and understand the application presentations and how you can ensure quality delivery and ensure that you're meeting those service levels, so that business can continue to take advantage of what that application provides to them.
So from an application perspective, you have both the cloud design piece that's referred to that, but, at the same time, you have to address the complexities of the application.
Verstraete: Tom, allow me to add one point. You talked about the application, but the next point associated with that is, on what device am I going to consume that application? Increasingly, we're seeing bring your own device (BYOD), and it's not just PCs, but also tablets, phones, and all of the other things.
We have to have the capability to manage those devices and make sure that we have the appropriate security levels and that they're compliant, so that I can run my enterprise applications on those devices without any trouble. That complements all of this.
Dana, to go back to a question that you had earlier, this is where all of these things are starting to come together. We talked about Cloud OS and the infrastructure and the environment, so that I could build on my applications. We talked about the Application Transformation Services, which allow us to put those applications on top of that. And we're talking about the other extreme, which is consuming those applications, and the devices on which we are starting to consume those applications.
Regardless of whether this is in a private cloud, a managed cloud, or a public cloud, that's where you start seeing the different parts and the different pieces coming together.
Why is the hybrid model so important with HP strategy?
Verstraete: Whether companies like it or not, most large enterprises today already have a hybrid model. Why? Because they have a lot of shadow IT, which is consumed outside the control of IT. It's consumed from external services, being in most of the cases public clouds. So that's already a fact of life.
Why is that used? Because there's a feeling from the business user that the CIO can't respond fast enough. So the CIO had better understand the potential issues related to the security and compliance of what is happening, and start acting on it.
He can't speed up his delivery of what the business is looking for by developing everything himself and taking the old fashioned approach. I choose an application. I test the application for six months. I install the application. I configure the application, and two years down the road, I deliver the application to the business users.
What becomes clear quickly to a lot of CIOs is that if they take a hard and cold look at their workloads, not all workloads are the same. Some of them are very specific to the core of what the enterprise is doing. Those should stay within their private cloud.
There are a bunch of other things that they need to deliver. Frankly, they are no different from what their competitors are doing. Do those need to be in a private cloud or could they be in another type of environment, a managed cloud, or public cloud? That automatically brings you to that hybrid environment that we're talking about.
New core competency
Paul Muller, how is hybrid perhaps the new core competency for IT, managing hybrid processes and hybrid systems and managing the continuum?
Muller: Again, Dana, you get to the core of the issue here, which is that it's about a shift. This is a generational shift in how we think about building, buying, and integrating IT services in the service of the business or the enterprise, depending on where you work.
It's about a couple of key shifts. It's about the balance of power shifting from IT to the business. We have probably said this countless times over the last three decades, but the simplicity, the focus on user experience, the ease with which competitive services can be procured from outside by laypeople from an IT standpoint has created a symmetry in the relationship between business and IT that no one can afford to ignore.
The second generational shift is the speed with which people expect response to their ideas. Techniques like agile and dev-ops are changing the way we think about building and delivering services.
Finally, to your point, it used to be that you either build or you buy, you either outsourced everything or you did it all yourself. Now, we live in a world where you can consistently do both. I don't believe that the majority of IT professionals are ready for that new reality in terms of processes and people, not to mention the software stack, the infrastructure stack, on which they're building services.
There's a lot of work to be done. It sounds daunting. The good news is that if you take a smart approach, some of the work that Tom and our Professional Services and Technologies Services team have been working on, it helps ease that transition and avoid people repeating the mistakes of some of the early adopters that we have seen.
Access to expertise
Tom Norton, as we factor what Paul said about transitioning the organization from supporting technology to supporting the continuum of a hybrid approach, how big a change is that for an organization?
Norton: It's a significant change, when you think of how traditional support structures have been. When you look at more complex systems, and you can think of a hybrid cloud environment as being a complex cloud system itself, traditionally support structures have been component-based and they've been infrastructure-based, or application-based. So you look at a storage support solution, or you may look at a network support solution or a compute solution itself.
When you start thinking of a complex system, like a cloud model, and especially a hybrid cloud model, where you have varying delivery mechanisms and varying supporting structures, supporting that can be a very complicated issue. It's one that many organizations are unprepared to do, especially if they're going to try to approach it strategically, as opposed to being a opportunistic-type cloud environment.
What IT is trying to do today and the question they keep asking is how they can view this as being that kind of ecosystem that has a singular support structure to it, where they can get access to expertise.
That's what HP is stepping up to do. With our own experience, across the spectrum, building on-premises and private, working in the managed infrastructure places, we have public cloud experience and we also have the experience of the integration across all of those.
The benefit from a services perspective for our customers is that we can help break down those isolated barriers in singular cloud services that a customer is consuming and give them a support structure that bridges all of those and truly approaches a converged support structure for managing that hybrid environment. That's what we're working towards, and that's where our announcements have been all about.
What is the most important change that HP has brought to the cloud landscape with this series of announcements?
Verstraete: Two things — first, and you hit the nail earlier, the whole concept of hybrid cloud, looking at multiple ways and multiple clouds to address the needs of the business. And second, within that frame of hybrid cloud, making sure that there is consistency across the different clouds, and that's where we're using OpenStack.
Paul Muller, what's different in your mind about what HP has been doing this week?
Muller: It is all about accelerating the introduction of applications and improving the user experience. It is not about technology for technology's sake. The single biggest difference.
And lastly, Tom, what jumps out at you as a differentiator in terms of the market in general and what HP is doing?
Norton: I think the market is looking for someone that can help with the integration component pieces of it. As the hybrid and heterogeneous deployments continue to grow and more and more services are offered that way, organizations need help consolidating that into a more integrated approach, so they have that kind of overall cloud concepts that give them the value they are looking for. So it's becoming more and more about integration.