What is Grid Computing? Why is it happening now? And what does Grid mean to enterprise customers and consumers? We recently had a chance to discuss these questions and more with Irving Wladawsky-Berger, Vice President, Server Group Technology and Strategy. Here's what he had to say.
Rayme: There are three times I can think of that IBM has turned itself on a dime to adopt an emerging technology. When the Internet began to take off, IBM mandated integration across the product portfolio. When Linux began to be used to make free Web file and print servers, again IBM made a corporate decision to adopt and support Linux across all e-servers, middleware and IBM Global Services. Now IBM’s new CEO, Sam Palmisano, has made it clear that IBM will support Grid Computing across the company.
In hindsight, the Internet and Linux decisions have certainly worked out well, but when the decisions were made, the outcome wasn’t so clear. Since you’ve been at the center of each of these decisions, this is a good opportunity to ask you, why these three? What do they all have in common that makes them worth steering an entire company the size of IBM directly at them?
Dr. Wladawsky-Berger: If you look at all these three areas: the Internet, Linux and now Grid Computing, they all started, grew and became widely adopted in the research communities. The Internet, of course, started much earlier. But let’s say by the early 1990s the research communities around the world were using it heavily, and they all started using the World Wide Web around this same time.
Linux got its start in the same way. Linus Torvalds and colleagues started building up Linux and then it started to be used more and more in universities, in the supercomputing communities, and in research labs. The story is similar with Grid Computing. Last year for example, IBM was involved in some very large Grid deals in supercomputing, in life sciences with universities, with governments, building up their supercomputing capabilities. All three -- the Internet, Linux and Grid Computing -- have the promise of making it big in the commercial world. Or let me say better... at a certain point in time each of these three looked like it was ready to break through from the research world into the wider commercial world, just as was the case with the Internet around 1995. It was the case with Linux around 1999, and it’s the case with Grid Computing now.
The key decision that IBM made as a business with the Internet and Linux, and is making now with Grid is to be at the vanguard of making these technologies break through to the commercial world. We do not want to sit back, wait until it happens, and then join the parade. You know, there are risks in being ahead of everybody because that means we’re making a bet that the Internet will be big in the commercial world, that Linux is good for our enterprise customers, that Grid Computing is going to have a huge presence in IT in general. But I think in each of these three cases, we made that huge leadership decision and we took the plunge, and in the first two we saw everybody else follow... some sooner, some later... and I believe with Grid Computing something similar will happen.
Rayme: The idea of an information utility, the idea of making enormous general purpose computing resources available anywhere seamlessly is quite a powerful idea, but it’s been around for a while. In fact, I think the first I heard of it was back in the ‘80s. Danny Hillis, as the co-founder of Thinking Machines, used to talk about it some, but the idea never took off. What’s different now?
Dr. Wladawsky-Berger: The idea of accessing computing and data without having to own the computers probably is even older than the 1980s, because at some level that’s what time-sharing computing was about, and that probably was back in the 1970s. It’s a very nice thought, which comes from watching other successful utilities. You know, people are used to the fact that they use electricity, they turn on water faucets, they use the telephone, they get access to television, and in none of these cases do they have to own the equipment that generates electricity. They don’t have to own the water supply, or the telephone switches, or the broadcasting companies. Everybody says, "Gee wouldn’t it be nice if I can similarly plug in to get access to applications and information?"
Now, to do this you need a few major ingredients that really haven’t been available in the past, but are becoming increasingly available now. Number one is bandwidth. If you’re going to access computing as a utility, you need a significant amount of bandwidth because otherwise you may access it, but the service that you get, the response time, is going to be significantly poorer than if you had the computer next to you, or at least under your control. The fact that bandwidth is becoming increasingly plentiful and increasingly inexpensive, with the promise of that trend continuing, is a major reason why computing utilities are thought of more seriously.
The second huge reason is standards. That is, if you want access to a service as a utility, you pretty much want everybody to use similar standards. We’ve seen that be very successful in the World Wide Web providing access to content. I know it’s hard to imagine what the world was like before the World Wide Web became widely used, but remember there was no common access to content. There were no browsers. There was no HTTP. There was no notion of Web sites, and therefore you couldn’t access anything. Then, because of the agreement on standards, all of a sudden the World Wide Web flourished and the rest is history.
We are now entering the stage where access to applications, access to data, access to computing power, access to storage is becoming standardized through efforts like Web services and Grid Computing, which are defining those standards, and I believe that having these commonly agreed upon standards is going to make it far easier for everybody to answer questions like "do I want to do it myself, or simply access it as a utility?"
Rayme: Increasingly, Web services are being proposed as a better way to deliver value to customers. Do you think all end user apps on the Grid will be delivered as Web services? And do you think cycles, bandwidth, data caching for virtual resource management and so on, will be themselves offered as Web services?
Dr. Wladawsky-Berger: Yes, and in fact last week at the Global Grid Forum, IBM, in collaboration with Globus, which is one of the main open community organizations for Grid Computing, submitted a specification called the Open Grid Services Architecture to do exactly that..– to merge Grid Computing protocols with Web service protocols and use Web services standards like SOAP, Simple Object Access Protocol; WSDL, which is the XML based standards for exchanging content; and other standards for services. So the two are definitely coming together, and that’s really good news for everybody.
Rayme: Okay. Let’s talk a bit more about the Web services piece. What will Web services really look like? There have been some examples that have been cited... financial transaction clearing comes to mind. But they're sometimes difficult to visualize. Do you think there’ll be, for example, flavors of a particular Web service for e-business to emphasize this or that capability, more or less, at this or that price point, as service providers try to distinguish themselves?
Dr. Wladawsky-Berger: Yes. Let me tell you, I think that Web services will be as rich and come in as many as flavors as civilization itself; as services in our civilized society. If you ask if there are special ways of dealing with lawyers and accountants and stores and hospitals, the answer is absolutely, but you share what you can amongst all of them. For example, in the US we generally speak in English. We use documents. We use credit cards, so there is a tremendous amount of sharing across all of those, but then there are unique aspects for each. So what you do in a hospital is likely to be somewhat different from what you do in buying milk in a supermarket, which is different from dealing with a lawyer and other types of relationships and transactions.
I think you’ll find similar things happening with Web services. There’ll be a common layer that pretty much all Web services will use, and then Web service providers will each start adding their own "vertical differentiation," depending on what industry it's for, what application it's for. And the reason is that eventually you want these Web services to become something that human beings are very comfortable using, and that human beings can adopt in their normal business as opposed to having human beings forced to change the way they think about their business or jobs in order to come down to the level of the computer.
Rayme: This all sounds a lot like the early days with the Web, when different Web sites offered different value, and companies would emerge with differentiating factors to set them apart from the competitors: Google and Yahoo, or Fatbrain and Amazon for example.
Dr. Wladawsky-Berger: Yes, I think that that’s a very good point, that there will be some differentiation that is because of the different industries, but then a lot of the innovation in the marketplace will be among different companies that are using the same standards. These companies will figure out how to make their web services available to people in particularly attractive ways, in particularly easy-to-use ways, in particularly flexible ways; and all that innovation is only to the good. That’s how things advance.
Rayme: You mentioned making these services available to people. The idea of the Grid as a compute analog to the electric utilities is tantalizing, but to make the analogy complete it needs to go beyond the business-to-business level and provide value to real consumers, to end users. And it has to do so seamlessly. Do you think that will happen? And if so, what do you think that might look like?
Dr. Wladawsky-Berger: Yes, I think it will happen, and I think that when it happens right, people won’t even know it’s happening. For example, I think that security is something that today is not well integrated into systems. That’s why often people don’t use security levels that are as sophisticated as they need to be. But I think that if you have security clearances that go to a totally different site to be analyzed by an application and come back integrated transparently into a normal application, all of a sudden people are using them because that security clearance is being processed in a very transparent manner. People don't even notice, but they are much more secure.
Another example might be better quality of service. Today when people go to a Web site, that Web site may be very busy, or may even be down. One of the promises of Grid is that the system would have multiple copies of Web sites, would have redundancy built-in. Then the system will make decisions to route you to that site that can most quickly handle your request. Now again, as a human being you could care less about all that, any more than you care whether your phone call is being switched through Kansas City or Chicago or some place in between, but the net result will be an Internet that performs much faster with much higher availability and much higher security.
Rayme: Okay. Well, this has been a great view into Grid, its abilities and what it offers e-business going forward. Is there anything else that you’d like to say to our audience?
Dr. Wladawsky-Berger: Yes, I would say that in the end you decide to embrace an emerging technology because it is the right thing to do for our customers. I remember last summer discussing with Sam the decision to embrace Grid, and one of the first things he said was that this is a value proposition we can explain to everyone very quickly so, let’s make sure we can actually build it and then if we can do it, let’s go for it.
Grid Computing is really the natural evolution of the Internet. This is really looking at the Internet with all it’s promise of universal connectivity and reach, and making it work far better by bringing the qualities of service that people are used to in enterprise computing, and the qualities of service that we all have gotten used to in utilities like electricity, telephone, our transportation system, all of which tend to be up all the time.
So the promise here is our customers will be able to have their cake and eat it too. They’ll be able to have all the openness, flexibility and reach of the Internet, but now we will add to that the great quality of service, the great efficiency and reliability that we have become famous for in enterprise computing.
Rayme: We appreciate your taking the time to talk to us.
Rayme Jernigan is the managing editor of the IBM Linux portals. He has published articles through several publications and content sites including JavaWorld, IBM/developerWorks, and Javasoft at Sun Microsystems. He was the founder and first president of the Triangle Java Users Group, and can be reached at rayme@us.ibm.com.
Note: All trademarks are the property of their respective holders.
