Wladawsky-Berger on the "state of e-business"

June 15, 2001

Irving Wladawsky-Berger, vice president of technology and strategy, IBM Server Group, addressed the e-business Convention and Expo in San Jose, Ca., Tuesday, June 12.

In my talk today I'd like to address the state of e-business, and offer some perspective on what we at IBM see happening. I'll talk about the drivers that we believe continue to push e-business forward, and then finish by examining the key challenges that we as an industry face in this e-business evolution that will continue, we believe, for many years to come.

History is a very good guide post to what has happened in the 18 months since I last keynoted e-business Expo. When we look at history we see that major transformational technologies typically go through a number of phases as they develop.

The evolution of technologies like electricity or the automobile or the telephone all display a dynamic similar to what has been happening with information technology and e-business over the last five years or so.

First, there is the period of discovery: Boy, isn't this wonderful? Isn't this fun? You can do incredible things. You can start a new business. After that, you can buy a Porsche Carrera, or maybe these days a BMW Z8. And a little later maybe an island in the Caribbean.

Sadly, that period doesn't last. At some point, it has to come to an end. But the technology continues to move forward into a much more serious era of hard work. That continues for many, many, many years, with improvement after improvement.

Later, with the passage of time, you look back and see how much the landscape has changed as a result of all those additive improvements. I believe we have definitely entered the second phase of e-business, the phase of serious e-business.

I say that, for one thing, because, despite the recent decline of the dot-coms, the number of people connected to the Internet continues to rise rapidly. So does the number of wireless phones, and the number of devices attached.

And when we look into the future, we see a huge number of new information appliances emerging.

So, for the foreseeable future we see more and more connected people and devices which have to be served.

We've also been saying for a while that a digital economy is being built, and in fact the key element of this digital economy is the business that businesses do with each other. It's a huge driver, and more and more enterprises engage in this business-to-business commerce, leveraging the Internet and leveraging e-business.

Every statistic we have shows the amount of B2B business going up over time -- sometimes faster and sometimes slower but all pointing very much along these same ascending lines.

So, between the numbers of people and devices connected and the volume of business that we see growing on the Web, there is every reason to believe that e-business is advancing, only in a more serious environment.

Nevertheless, to better understand what's going on, we commissioned a study of e-business adoption around the world. The study looked at seven countries: the US, Canada, the UK, Germany, France Italy, and Japan.

We interviewed 33,000 companies, large and small. We wanted a statistically significant study that truly represented the marketplace.

We grouped the results in three major categories: first, early adopters -- organizations that have started to use e-business for publishing information and doing simple transactions -- then those which had integrated, had started to bring together their various processes and applications.

And finally, we looked at those businesses that had entered the more advanced stage in which they actually were transforming their businesses, deciding which processes to keep in-house and which to outsource, while keeping everything integrated.

We found that over 90 percent of those surveyed around the world had adopted e-business in one form or another. The precise number was 93 percent.

This included some very small businesses. But, as you can imagine, the larger the business the greater the probability that it had adopted e-business. The smaller the business, the less extensive the adoption of e-business.

We found, not surprisingly since we are early in the evolution of e-business, that the vast majority -- about 80 percent -- are still using the Internet for publishing information or for doing simple transactions.

That means that there is a huge potential for the growth of e-business around the world, and for companies to embrace more and more technology for more and more of what they do.

We found that about 15 percent of those surveyed had gone on to the next phase, that is, to integration internally and externally. And about five percent had started the march to truly transforming themselves and moving toward what we call "Dynamic e-business."

(Actually, for large businesses, it was hard to characterize the total organization, since some elements of the business may still be in early stages whereas other processes may be in the more advanced stages).

Overall, the results are a very good indicator of, number one, the fact that e-business is real; over 90 percent of businesses use it. And number two, there is a lot more to be done over many years in continuing to bring technology to businesses all over the world.

But one can still ask why, in spite of all the financial turbulence of the last 18 months, why is e-business still going forward?

We see two reasons. The first is almost obvious, but sometimes it's good to address the obvious. This stuff has real business value.

Companies adopt e-business because it helps them do what they do better. It helps them save money, it helps them better support their customers, it helps them better integrate their organization, and it helps them better connect to their whole value chain.

The best way to see this is to look at real e-businesses from among the many we have been working with across three broad spectrums -- the first one being companies that are using e-business to reach out to their customers.

I know that conventional wisdom insists that business-to-consumer e-commerce is dead -- that there is no money in B2C, that a lot of companies that started in B2C went under, that it's over.

Well, maybe it's dead if you don't have a good business model. But what we found is that if you are a company with a solid business, B2C is very much alive and very valuable. Let me give you a couple of examples.

One company that we worked with is Wolferman's in Kansas City which sells specialty baked goods. Wolferman's has been in business over 100 years.

Several years ago they started a catalog business to reach out to many more people. And then in the last couple of years they went online and put up a Web site. So here is Wolferman's selling baked goods over the Internet from Kansas City -- not generally considered a Mecca of high technology -- being very happy and successful on the Internet.

Why are they happy? Well, they've found that the cost of a transaction went from $2.70 when people call over the phone to order from their catalog to 45 cents when they do it online -- an 80 percent reduction in costs.

They found that they were already doing last year 10 percent of their catalog business online and that number will rise to 25 percent next year, which means that more and more of that business is being done at lower cost.

And they also found that they could reach a whole new audience: younger customers who never ordered from them before.

Another interesting example is Holt's Cigar in Philadelphia, also over 100 years old. They're selling $6 million worth of cigars a year online.

And that clearly doesn't include Cuban cigars, which, it's still illegal to sell. If that ever changes, who knows how much more they can sell online.

My favorite in B2C is a company called the Lehigh Valley Safety Supply Company which we started to work with at the dawn of e-business. They sell safety shoes -- the kind of shoes with a steel cap in the toe to protect workers.

It's a small company, it operates in the northeast -- New York State, Pennsylvania, Virginia, Maryland and others. They were doing business off a truck that goes to where the workers are and sells on-site.

They put up a Web site, and initially offered only part of their inventory over the Web. They were so successful that now they are selling all their products over the Web.

They are finding that over 90 percent of the sales come from outside their local territory in the northeast which they service physically. So they have significantly expanded their business by selling safety shoes over the Internet.

B2C may not be right if you don't know what you're doing, but boy if you have a good business, you've been around for a while, and you aspire to move your business into the future, B2C continues can be very successful.

Let me talk about another aspect of e-business that also seems to be doing quite well, that is, how companies are integrating their processes and using e-business to give all their employees and all their business partners access to information and applications.

A very good example is SNCF -- Societe Nationale de Chemin de Fer -- which is the French national railroad company that also oversees other forms of transportation.

They service about 800 million travelers a year and they ship about 135 million tons of goods across various countries in Europe.

They decided that to better control their business, control their expenses, keep track of shipments, they needed to build a gigantic data warehouse. And of course, they gave Internet access -- intranet access, in this case -- to all of their people who needed it. Just about all of their processes now connect to this, this major portal, with all its attendant efficiencies.

Goodyear is a similar story. In North America, Goodyear has 19 manufacturing sites and 5,000 dealers. They are integrating all of them so that a dealer can go to the Web and know precisely what tires cost this particular week, what promotions are underway, what's shipping, etc.

When they want to order a tire they don't have in inventory, they know how quickly they can get it because all of the inventory is online. They place the order online, they track the order online. Costs come down, dealer and customer satisfaction goes up. This is a very, very solid example of how e-business continues to move forward.

The last area I want to briefly talk about is the world of business-to-business commerce -- B2B -- and e-marketplaces.

One interesting example is a company in Europe, Freight Trader. Freight Trader is essentially a marketplace that brings together shippers with carriers.

They estimate that about 27 percent of trucks in Europe are empty at any given time, because they've made a delivery and they have nothing to take back on their return. What do you do when you have these trucks on the road empty?

You put that availability online. Over the Web, you let people know you have a truck going, let's say, from Paris to Barcelona, the time it's going, and the amount of space it has. Again, online you hold reverse auctions. Carriers bid and match their trucks with cargo.

As a result, last year they effected 114 million euros worth of transactions between shippers and carriers. And this is a very good example of a business that didn't exist before. Now trucking is more efficient, and the percentage of trucks empty is going down as a result of bringing all of the community together online.

So e-business continues to have value for businesses today and absolutely continues to offer even greater value into the future.

The second key driver of e-business is technology. It's important to remember that IT is information technology and that IT definitely depends on the continued advance of technology in order to keep moving into the future.

Every time we look at good old Moore's Law for guidance about the future in technology, we find that not only is Moore's Law alive and well, but in a number of cases its effects are growing even more dramatic.

You can see it daily in the newspaper. Last week it was strained silicon, a technique that IBM invented to get more performance out of silicon. Then we saw that Intel had developed new techniques for further miniaturizing circuits that they hoped can get microprocessors to 20 gigahertz over the next several years.

If you look at storage technology, what's going on in that world is almost scary, with capacities doubling every year or so with increasing price/performance. Look at what's going on in just about any aspect of technology from micros to storage to memory to fiber optic bandwidth, wireless bandwidth, to everything ... and you see that technology continues to move up, price/performance continues to improve radically and drive down prices precipitously. As a result people can buy a lot more technology.

The net effect of all this plentiful and increasingly powerful and inexpensive technology is that the e-business infrastructure continues to explode. It continues to add bigger and bigger servers, some doing transactions, some managing content, some doing supercomputing.

It continues to support many more client devices, as we observed at the beginning, with huge numbers of new information appliances expected to show up in the next several years.

The network itself is becoming far more intelligent with servers of all sorts infiltrating themselves into the network handling security, directory, caching, media-streaming and God knows what else.

And in fact, it's not too difficult to do a few projections and convince yourself that in five years a typical e-business infrastructure will have 10- times more technology than it has today. That's less than 60 percent growth a year, certainly within the bounds of what we are seeing in price/performance improvements.

So the technology is a huge driver, enabling many more kinds of computer devices which in turn enable many more kinds of e-business applications.

Is this an unmixed blessing or are there challenges ahead for us as a community and for vendors like IBM specifically? Let me show you one of our recent e-business commercials that points them up very clearly. Can you please roll the video?

[VIDEO VIGNETTE -- "The Deal" 60 sec.]

I didn't mean to do anything so crass as to show you one of our ads but, the story in the ad is a very good illustration of the challenges that we as an industry face in this technology explosion with more and more applications coming online.

And I'd like to turn now to three key aspects of those challenges, starting with the need to make our systems much more self managing.

When we in IBM talk to our customers, they almost always tell us that their number one problem today is managing their infrastructure -- providing the needed scalability, response time, availability and security -- all at an affordable cost.

Then we have to look them in the eye and say, "Boy, we know this is tough but we have some news for you." As a result of work in our laboratories and in those of our colleagues across the industry, five years from now your problem will be 10-times greater. That may be enough to maybe drive some of them out into a kinder, gentler business.

But seriously, management is a major issue, one that we need to do something about. The answer is to leverage a lot of that great inexpensive and powerful technology to make the systems manage themselves to a greater degree. After all, if these computers are so good ... let them be good not just at the applications, but at managing themselves, at handling as much as possible the optimization, the configuration, the protection, the healing that you would expect from any form of intelligence.

That, we believe, is one of the major, major areas where we as an industry need to make a serious investment.

There has been a lot of work. We clearly aren't starting from zero; there is a lot of stuff out there across systems from many vendors.

For example, we at IBM have been investing for quite a while in making our systems much more automated in their management. In fact, that is what has made mainframes, mainframes.

And when we made our most recent announcement of the new mainframe line, the z900, a lot of these new self-management capabilities were included: much more intelligent dynamic workload management, for example, permitting different workloads to grow and shrink depending on policies that the systems managers had articulated.

The system can take resources from lower priority applications and give them to higher priority applications. So it's an area where there is a lot of work but clearly given the prospects in the future we need to do a lot more.

That's why we are working in our research labs in this whole area of self managing systems. One effort is called project Oceano.

We see a world in which inexpensive technologies, inexpensive processors that you can rack and stack, are clustered, with thousands going to tens of thousands of processors. With the 1U form factors, we have a whole world of denser and denser servers coming along with better price/performance.

And there are applications waiting in the wings to use these server, from supercomputing applications to life sciences applications to e-utilities. By e-utilities, I mean gigantic server farms able to host many, many different applications for a variety of customers.

A problem in hosting so far -- a problem that Oceano is addressing -- is the fact that often different customers in the farm are rigidly segregated by a chain link fence, figuratively speaking. That may guarantee security, but it is horribly inefficient because, if you experience a surge in traffic as a result of a TV ad perhaps or a good review of your site, you have to get more and more physical processors inside that fence, which takes time and is expensive.

A better approach is to have virtual fences, partitions that absolutely guarantee segregation of the different applications, but permit the automatic addition of processors in real time and bring up all the software in real time, spreading the work now over say 150 processors instead of 50.

Then when traffic subsides you can start returning processors, without the data, of course so security is not compromised.

That's the kind of thing Project Oceano is designed to do. It's all a matter of micro code and software. But projects like this are going to make it much easier to put together, manage and deliver much higher service to customers.

Another important challenge is to provide self-healing technologies. Our mainframes already support very long mean-times to failure measured in decades.

Among the many reasons is that our mainframes have something called Instruction Retry, which means that, if an instruction fails, the system automatically tries again.

If it still fails, the system can resort to spare processors, move the state from the one that failed to the spare, bring it online, and send a message to IBM saying, "Bring me another spare because I had to use one of them." All the while, the system continues to work.

Until recently only higher-performance processors could afford that level of fault tolerance. As we look at increased density in microprocessors, as essentially we have more circuits in our micros than we know what to do with, we are learning what to do with the circuitry -- namely, make sure the system never fails. Add enough redundancy, enough spare capacity and enough intelligence so that the system just doesn't fail.

In fact, the more we've looked at this whole world of self-managing systems the more interesting it becomes. You don't have to look much farther than biological organisms that have survived for millions and millions of years to be inspired by how nature has handled the problems of making sure these organisms, from lowly lizards to human beings, can survive even though predators abound.

The system exists in a hostile world, God knows what's going to happen next. The answer is, more and more and more of the circuitry is devoted to staying alive, to making sure that you move on and keep going in spite of being surrounded by all kinds of hackers that will try to bring you down, denial of service attacks that will come after you, availability glitches, software updates that don't work -- and on and on and on.

I mean, the number of things that can happen to a poor computer which is out in the open, hostile expanses of the Internet is very large. Fortunately, the right research can help us make those computers better able to take care of themselves by building the necessary software and the necessary intelligence into them.

A second major challenge is how do you integrate all of the pieces in this exploding infrastructure with maximum flexibility?

The Internet showed us how much more valuable a system is when all the pieces work together, when from your PC you can access every application and every Web site out there.

And as the infrastructure keeps growing, connectivity and integration are more important than ever. There is only one answer to achieve this kind of connectivity: We must embrace standards that work across every system from every vendor on any architecture, and build software that embraces those standards, so that regardless of whose system it is all of the pieces go together.

This realization of the importance of standards, and open source software, is why IBM and others have so aggressively embraced Linux in the last few years.

Linux is more than just an operating system. Everybody knows it's built by a community, but it's important to reflect on the fact that this is a community of the top computer scientists in the world. This is a community where getting your design accepted in the next version of Linux is considered almost as positive as a physicist getting his or her paper published in one of the top physical journals. This is a big meritocracy.

Not surprisingly, as a result of that, Linux has come out with excellent features. It's very modular, it's very elegant, it works very well. And one of the best proof points of its modularity and elegance is that Linux is the only operating system -- I believe the only one that has become popular -- that runs on every single architecture.

It runs on Intel 32-bit and 64-bit. It runs on RISC processors -- our own PowerPC and everybody else's. It runs on IBM mainframes as virtual partitions. It runs on network processors. It runs on appliances out there like Tivo and many others that are coming along.

In fact, Linux is the only operating system that I can safely predict will run on microprocessors that have not yet been invented, let alone designed. Now, that's not bad for an operating system and as a result of its ability to provide a common operating environment on everything from an appliance to a mainframe, it makes the life of application developers and the life of IT personnel who have to move applications around much, much simpler.

That's why Linux is very popular because it brings value to a very important area: getting maximum application flexibility in an increasingly diverse world?

Another very important set of emerging standards is the whole world of Web services. We have built a great World Wide Web and standards like HTML and HTTP and DNS have helped make it happen.

But so far we have lacked the right standards to help applications connect with each other. Connecting and accessing content? We've done very well. But integrating applications like a CRM application or a supply chain application with any of the many e-business applications that you want to connect still requires too much human labor, too much programming labor.

We think that will start changing because just about the whole industry has embraced standards that we call Web services: the world of UDDI -- Universal Description Discovery and Integration for directories that help you find applications; SOAP -- Simple Object Access Protocol for sending messages around so applications can communicate with each other; and finally, WSDL which is based on XML for applications that exchange content.

Products are emerging. IBM just announced a few weeks ago the new version of WebSphere which totally embodies these new Web services interfaces. Just about every vendor is announcing initiatives supporting Web services, and we hope that, over the years, Web services will do for building an application Web what the original Internet standards have done for building the World Wide Web.

The final challenge that I want to discuss briefly and close on is access to expertise. Skills are hard to find. This stuff is not easy. We already see services being the largest portion of the business for companies like IBM and in fact for the whole industry. On this slide, you see estimates of the split of IT revenue across services, hardware and software.

We believe that over the next few years many more services will be available to businesses everywhere. So businesses will not hire the people themselves and do the work themselves, but increasingly will source everything from strategic consulting to Web hosting to outsourcing to training and on and on and on...

...not unlike what happens in civilization in general where you can set up a business and hire lots of the services that you need to support your business -- from cafeteria workers to maintenance people to accountants, lawyers and all the others who are not in your core business.

Essentially you have this virtual business where you perform your core function and you're surrounded by services from everybody else.

We see very substantial growth in the ability to provide these services. Standards like Web services will help a lot since it will make it easier to create services that everybody can use. And we think this will drive e-business into the future to many, many more businesses especially smaller businesses.

To conclude, e-business is alive and well. The hype stage may have ended, but that's how it has to be if you go on to the next stage of hard, serious work.

We are in that next stage. We have a lot of work to do. We have incredible technology propelling us forward. We have many new applications. We have many new possibilities. And it's going to be very interesting to look out into the future and see how information technology will continue to change the landscape of business.

Thank you very much.

Copyright 2001