Hybrid Service Has Its Head in the Cloud but Keeps Its Feet on the Ground
Let's discuss the specific concepts around converged cloud, information clouds, and hybrid services delivery.
To gain deeper insights on this topic, we're joined by Paul Muller, the chief software evangelist at HP; and Christian Verstraete, chief technologist for cloud strategy at HP. The discussion is moderated by Dana Gardner, principal analyst at Interarbor Solutions.
Listen to the podcast (24:44 minutes).
Here are some excerpts:
Dana Gardner: You've separated the notion of hybrid computing and hybrid delivery. Can you help me understand better why they're different, and what HP means by hybrid delivery?
Christian Verstraete: Hybrid computing typically is combining private and public clouds. We feel that many of our customers still have a traditional environment, and that traditional environment will not go away anytime soon. However, they're actually looking at combining that traditional environment, the data that's in that traditional environment and some of the functionality that's out there, with the public cloud and the private cloud.
The whole concept of hybrid delivery is tying that together. It goes beyond hybrid computing or hybrid cloud. It adds the whole dimension of the traditional environment. And, to our mind, the traditional environment isn't going to go away anytime soon.
Gardner: Paul, how has the traditional understanding of cloud computing as segments of infrastructure services changed?
Paul Muller: From that perspective, the converged cloud is really about three things for us. The first is having greater levels of choice. The key point that Christian just made is that you can't afford to live in the world of, "It's just public; it's just private; or I can ignore my traditional investments and infrastructure." Choice is critical, choice in terms of platform and application.
The second thing, though, is that in order to get great choices, you need consistency as an underlying platform to ensure that you're able to scale your people, your processes, and more importantly, your investments across those different environments.
The last one is probably the biggest area of passion for me -- confidence. We spoke a little bit earlier about how so many clients, as they move to cloud, are concerned about the arm's-length relationship they have with that provider. How can I get back the confidence in security and service levels, and make sure that that confidence is consistent across both my on-premises and-off premises environments?
Verstraete: People have started looking at cloud from pure infrastructure, reuse, and putting workflows in some particular places in infrastructure. The world is moving beyond that at the moment. On one end, you have Software as a Service (SaaS) starting to play and getting integrated in a complete cloud environment and a complete cloud function.
We also have to realize that, in 2011, the world created about 1.8 zettabytes of data, and that data has a heck of a lot of information that enterprises actually need. And as enterprises understand what they can get out of the data, they want that data right there at their fingertips. What makes it even more interesting is that 90 percent of that data is unstructured.
We've been working for the last 30 years with structured data. We know all about databases and everything, but we have no clue about unstructured data. How do I know the sentiments that people have compared to my brand, my business, my product? That's the sort of question that's becoming important, because if you want to do warranty management or anything else, you want to understand how your users feel. Hence, the importance of all of this data.
Muller: I'd add something else. We were here with the Customer Advisory Board. We had a pre-meeting prior to the actual conference, and one of them said something I thought was kind of interesting, remarkable actually.
He said, "If I think back 30 years, my chief concern was making sure the infrastructure was functioning as we expected it to. As I moved forward, my focus was on differentiating applications." He said, "Now that I'm moving more and more of the first two into the cloud, my focus really needs to be on harnessing the information and insight. That's got to become the core competency and priority of my team."
Verstraete: There's one element to add, and that is the end-user. When you start talking about converged clouds -- we're not there yet, but we're getting there -- it's really about having one, single user experience. Your end-user doesn't need to know that this function runs in a public cloud, that function runs in a private cloud, or that function runs in the traditional environment.
No. He just wants to get there and use whatever it is. It's up to IT to define where they put it, but he or she just wants to have to go one way, with one approach -- and that's where you get this concept of a unique user experience. In converged cloud that's absolutely critical.
Gardner: Another term that was a bit fresh for me here was this notion of composite hybrid applications. This was brought up by Biri Singh in his discussion. It sounds as if more and more combinations of SaaS, on-premises, virtualized, physical, and applications need to come together. In addition to that, we're going to be seeing systems of record moving to some variety of cloud or combination of cloud resources.
The question then is how can we get to the data within all of those applications to create those business processes that need to cut across them? Is that what you're talking about with Autonomy and IDOL? Is that the capability we are really moving toward, combining data and information from a variety of sources, but in a productive and useful way?
Verstraete: Absolutely. You got it spot on, Dana. It's really about using all of the information sources that you have. It's using your own private information sources, but combining them with the public information sources. Don't forget about those. Out of that, it's gathering the information that's relevant to the particular thing that you're trying to achieve, be it compliance, understanding how people think about you, or anything else.
The result is one piece of information, but it may come from multiple sources, and you need an environment that pulls all of that data and gets at that data in a useful form, so you can start doing the analysis and then portraying the information, as you said, in a way that is useful for you. That's what IDOL and Autonomy does for us in this environment.
Muller: This has to be not yesterday, not today, but in real-time. One of the critical elements to that is being able to access that information in real-time. All of us are active in social media, and that literally reflects your customer's attitudes from minute to minute.
Let me give you a use-case of how the two come together. Imagine that you have a customer on a phone call with a customer service operator. You could use Autonomy technology to detect, for example, the sound of their voice, which indicates that they're stressed or that they're not happy.
You can flag that and then very quickly go out to your real-time structured systems and ask, "How much of an investment has this client made in us? Are they are high net worth customer to us or are they a first-time transactor? Are they active in the social media environment? What are they saying about us right now?"
If the pattern is one that may be disadvantageous to the company, you can flag that very quickly and say, "We want to escalate this really quickly to a manager to take control of the situation, because maybe that particular customer service rep needs some coaching or needs some help." Again, not in a week's time, not in a month's time, but right there, right now. That's a really important point.
Gardner: This is a good vision, but if I am a developer, a business analyst, or a leader in a company and I want a dashboard that gets me this information, how do we take this fire hose of information and make it manageable and actionable?
Verstraete: There are two different elements in this. The first thing is that we're using IDOL 10, which is basically the combination, on one hand, of Autonomy and, on the other hand, of Vertica. Autonomy is for unstructured data, and Vertica for structured data, so you get the two coming together.
We're using that as the backbone for gathering and analyzing the whole of that information. We've made available to developers a number of APIs, so that they can tap into this in real-time, as Paul said, and then start using that information and doing whatever they want with it.
Obviously, Autonomy and Vertica will give you the appropriate information, the sentiment, and the human information, as we talked about. Now, it's up to you to decide what you want to do with that, what you want to do with the signals that you receive. And that's what the developer can do in real-time, at the moment.
Gardner: Paul, any thoughts in making this fire hose of data actionable?
Muller: Just one simple thought, which is meaning. The great challenge is not lack of data or information, but it's the sheer volume as you pointed out, when a developer thinks about taking all of the information that's available. A simple Google query or a Bing query will yield hundreds, even millions of results. Type in the words "Great Lakes," and what are you going to get back? You'll get all sorts of information about lakes.
But if you're looking, for example, for information about depth of lakes, where the lakes are, where are lakes with holiday destinations, it's the meaning of the query that's going to help you reduce that information and help you sort the wheat from the chaff. It's meaning that's going to help developers be more effective, and that's one of the reasons why we focus so heavily on that with IDOL 10.