TMF Interview With Inktomi Corp. Chairman, President, and CEO David Peterschmidt
With Brian Graney (TMF Panic)

October 8, 1998

Inktomi Corp. <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: INKT)") else Response.Write("(Nasdaq: INKT)") end if %>, based in San Mateo, California, is a developer of scalable software for the Internet. The company's search technology is behind several of the Web's most popular search engines, including Lycos' <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: LCOS)") else Response.Write("(Nasdaq: LCOS)") end if %> recently acquired HotBot and Microsoft's <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: MSFT)") else Response.Write("(Nasdaq: MSFT)") end if %> MSN Internet Search. The firm also develops Internet caching software and recently entered the e-commerce field through its purchase of online comparison shopping software firm C2B. We caught up with Inktomi's chairman, president, and CEO David Peterschmidt, who was formerly the COO at database software vendor Sybase Inc. <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: SYBS)") else Response.Write("(Nasdaq: SYBS)") end if %>.

TMF: What you are trying to do at Inktomi?

"The fundamental business and vision of Inktomi is that of scaling the Internet. What we mean by that is, we're providing the software infrastructure of the Internet."

Peterschmidt: The fundamental business and vision of Inktomi is that of scaling the Internet. What we mean by that is, we're providing the software infrastructure of the Internet. As it gets larger, both in the amount of documents that are on the Internet and the number of users trying to access information on the Internet, [scaling allows] the Internet itself to become a faster medium as it gets bigger. So in essence, [we want to] scale up its capability to handle more and more growth.

To put that into perspective as to what kind of challenge that is, it's estimated that we have somewhere around 330 million to 350 million documents on the Internet today, and that we're adding a million documents a day. By 2002, it's projected we will have over 7 billion documents. We know that we are attracting millions and millions of users that want to get access to that information, so computing has to get faster and the whole medium has to get faster. Generally, the way that happens in big architectures is through software. Hardware, transmission, and all that is going to help, but software becomes one of the primary ways you facilitate speed and access.

TMF: What is Inktomi doing to establish itself as the leader in the architecture software industry?

Peterschmidt: We started out trying to solve the first big problem of the Internet, and that is: "If there are that many millions of documents out there, how do I find what I want?" We developed the first of the big, scalable search engine applications. So we provide the power behind a lot of the most prominent search locations on the Internet. And those are companies like Yahoo! <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: YHOO)") else Response.Write("(Nasdaq: YHOO)") end if %>; the new Microsoft <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: MSFT)") else Response.Write("(Nasdaq: MSFT)") end if %> Internet system MSN, which is powered by Inktomi; HotBot, which has probably been recognized as the most sophisticated search site on the Internet, which Lycos <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: LCOS)") else Response.Write("(Nasdaq: LCOS)") end if %> just bought; CNET <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: CNWK)") else Response.Write("(Nasdaq: CNWK)") end if %>; the new Disney <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: DIS)") else Response.Write("(NYSE: DIS)") end if %> Internet guide. We're establishing our leadership position through the types of partners that we have using this technology. We think we've captured the dominant players today on the Internet with the exception of one or two. And we believe, over time, just about all the dominant sites will use that search application.

The second problem that we solved is: "Once I find the document through the search application, how do I then get the document to my desktop and not sit there watching the stars go by on my Netscape <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: NSCP)") else Response.Write("(Nasdaq: NSCP)") end if %> browser for five minutes?" The second piece of the equation is what we call network caching, and the product is called Traffic Server. The premise behind that is that about 40% to 60% of all the traffic on the Internet is redundant content requests. That is, we're asking for the same information to be sent to the same location in fractional periods of time.

A good example we give is the BankAmerica building here in San Francisco, [which] probably houses 10,000 people. You can just imagine how many times the front page of the Wall Street Journal is electronically transferred from the server in New York to the BankAmerica building in San Francisco in the first 30 minutes of any working day. The caches actually will trap that request the second time it is asked for. And every time after that, rather than travelling electronically to New York, everybody else will get that out of the cache. That's the second piece of the application scenario that we see to help the Internet get faster.

Every user of America Online <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: AOL)") else Response.Write("(NYSE: AOL)") end if %> goes through Inktomi's network cache product. In fact, AOL estimates that 1.2 billion hits a day go through that. How does that help the user? Probably the most recent example was when the Starr report was released, and everyone that went to AOL was amazed because they didn't wait to get the report. That's because the network cache system is fully deployed at AOL. In fact, AOL estimates that no one waited longer than 1.5 seconds to get the Starr report.

TMF: Do you think that's really going to be where the profits are going to come from down the road -- from the caching -- and we're going to move away from the emphasis on the search?

"... [T]he problem of searching [on the Internet] is going to get more sophisticated and more complex over time."

Peterschmidt: No, I think search is always going to be there. If you think about the fact that we're going to have 7 billion documents out there, the problem of searching is going to get more sophisticated and more complex over time. The search itself, as far as the user is concerned, has got to become simpler and simpler all the way down the road. So it's going to be taken for granted. But the search business for us is exploding right now. We have [many] requests coming into the company from organizations that want to develop a search capability for their site. The network cache will be a big market. It's already estimated that it's going to be an over $2.2 billion market just over the course of the next few years.

Now the third application that Inktomi has gotten involved in is the next big wave of activity on the Internet, [which] is going to be shopping, and comparative shopping especially. We just recently acquired a company called C2B, which is about consumer-to-business shopping. And we believe in the next year now, we're going to start to see the evolution of major moves in this area on the Internet. Amazon.com <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: AMZN)") else Response.Write("(Nasdaq: AMZN)") end if %> just bought a company called Junglee and Excite <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: XCIT)") else Response.Write("(Nasdaq: XCIT)") end if %> has bought a shopping product company, so you'll see [more of] that.

Underlying all of these applications that Inktomi has, is that they sit on top of the core technology of Inktomi. And that core technology is what really sets Inktomi apart as the leading infrastructure provider to the Internet. The core technology allows us to take off-the-shelf work stations and tightly couple-cluster them together so that they operate as one virtual computer. Now, the benefit of that is that it gives almost linear scalability to growth in computing power, but you get to do it one work station at a time rather than using big, heavy supercomputers. That allows each application, no matter how many users or how many documents or how many transactions, to keep growing with the Internet without diminishing the speed and response that occurs. To give you an example, today we have over 200 Sun <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: SUNW)") else Response.Write("(Nasdaq: SUNW)") end if %> Ultra II's in a cluster, operating as one virtual machine. And each of those Ultra II's has 200 or 300 MHz CPU cards in them. So there are 400 computers operating as one virtual machine.

TMF: Do any of your competitors have anything close to that kind of scale?

Peterschmidt: No, and that is why Inktomi has been so successful. It's very difficult to do what we've done, and we're actually the outgrowth of three years of government-funded research at Berkeley. This is about a five-year barrier to entry to do this because we've got the benefit of the three years of government research, and we've put another two-and-a-half years of commercial development on top of that government experience.

TMF: What other companies could you fairly compare Inktomi to, in terms of where you're at in your growth strategy or what you're trying to do in the Internet field.

"Wall Street tends to be looking at Inktomi now in the same way that it looks at Yahoo! or AOL. It assumes that Inktomi is one of these defining companies."

Peterschmidt: Let's take it at two levels. Let's take it first at the Wall Street level, or "How does Wall Street look at Inktomi versus other companies?" Wall Street tends to be looking at Inktomi now in the same way that it looks at Yahoo! or AOL. It assumes that Inktomi is one of these defining companies. If you look at stock prices, AOL, Yahoo! and Inktomi move almost in lockstep and unison up and down in the marketplace. So the fortunes of these companies seem to be tied together, at least so far as Wall Street is concerned. And Wall Street has already given Inktomi well in excess of a $1 billion market valuation, even though the company just went public four months ago.

But the other way you would look at Inktomi is that there is no direct company that competes with Inktomi. In fact, there is no like-kind company today. Who Inktomi's competitors would be would depend on which segment you're in. In the search application, we used to see Alta Vista as a potential competitor. But they seem to have gone more and more to a branded strategy than an OEM strategy. They are also the only company that has any shot at scalability of a large size. But quite frankly, from what we know about them, they use about seven times the amount of hardware to achieve the same computing power. So over time, they probably won't be economically viable as a competitor to us.

In the network cache business, Cisco <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: CSCO)") else Response.Write("(Nasdaq: CSCO)") end if %>, Network Appliance <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: NTAP)") else Response.Write("(Nasdaq: NTAP)") end if %>, and a company called CacheFlow all have what they call network caches. However, they do that with an appliance approach. [That] means they take a piece of hardware, have a closed proprietary operating system on that hardware, and then embed a cache application. You buy it as a box. Our approach is an open software architecture. You don't buy the hardware from us, you just buy the software application. And you can run that software on DEC hardware, on Sun hardware, on Silicon Graphics <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: SGI)") else Response.Write("(NYSE: SGI)") end if %> hardware, and shortly on Microsoft NT hardware.

By being an open software environment, we see caching not as an appliance, but as the launching platform from which you launch value-added services in the network. So open software becomes important. An example of value-added services in a network from the cache is [our] recently announced agreement with RealNetworks <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: RNWK)") else Response.Write("(Nasdaq: RNWK)") end if %>, where their video player is being engineered to be integrated into the cache. So the cache puts content out at the edges of the network, and you'll then be able to launch video services out of the cache close to where the user is.

Another example of value-added services on the cache is what's called transformation. You cannot, today, view a Web page from a palm-top device. [But if] you put a transformation module on top of the cache, the cache will actually reformat the page on the fly depending on what the requesting device looks like. We think that approach is going to win over the appliance approach. And we're the only ones with an open software architecture for network caching and the scalability. So, we think we're in good shape. We believe we have well in excess of 50% market share around the world today in that product and continue to innovate there.

In shopping, I believe the competitors will be companies like Amazon.com. They have said that they're clearly going to be in the shopping business and that they're going to try and provide a shopping service. We would provide our service to the other major portals. We would like to win and convince companies like AOL and Microsoft, who are our partners today, to adopt our shopping application. So that's kind of the picture of these applications and who might compete with them.

TMF: Finally, what other companies working in the Internet sector, on a personal basis, do you admire the most?

Peterschmidt: I admire Yahoo! a great deal. They have a very gifted management team. They've got very bright, young innovators coupled with a very mature, very sharp management team. We respect them a great deal. We think that what the guys at RealNetworks are doing is tremendous. They're fighting right into the teeth of Microsoft and they're doing a great job with, again, very good technology and a great vision. AOL, obviously, we respect a great deal. We see those companies as very strong companies and we like what they're doing.

TMF: I'd like to thank you very much for your insights today.

Peterschmidt: Thank you for the opportunity. We appreciate it.

Related links:
Inktomi message board
Inktomi website

Got an idea for StockTalk? Who would you want us to interview and what should we ask? Drop us a note at [email protected].

Transcript Archives