to homepage
 Weekly emails: how to advanced search
 Glossary lookup:


Loosely Coupled weblog

Friday, August 19, 2005

Why Intel bought Sarvega

Just two months after Cisco's AON launch, now Intel has waded into the XML appliance market. The company is keeping tight-lipped about why exactly it has acquired Sarvega, but there are various pieces of evidence that point to the likely reasons. One thing is clear: Intel is just as determined as Cisco to be an influential player in this emerging and potentially very significant market, which regular readers may recall we analyzed in some detail in the latest issue of the Loosely Coupled monthly digest.

This is not Intel's first foray into XML appliances. Five years ago, the company launched an appliance called the XML Director, designed to route traffic based on the content of XML messages. A fine idea, but way too early. “It was so far ahead of its time that there was no XML to direct,” explains Mark O'Neill, CTO of network security vendor Vordel, whose company, which has Intel as one of its investors, contributed to the project.

The work that Intel had done on the XML Director appliance was spun out into Tarari, a startup established in 2002 to design and market XML acceleration chipsets. Tarari's chips are used in a number of vendors' products, including Cisco AON. A year or so later, Intel invested in Sarvega, which is one of the few XML appliance vendors that uses general-purpose hardware platforms rather than incorporating dedicated accelerator chips.

If Intel already had an investment in Sarvega, why has it now bought the company? One possible explanation might have been that Sarvega was in financial difficulty, but Sarvega's management has denied this. Competitors say that Sarvega was not making much headway in the market, but the company, founded in 2000, had at least $20 million venture funding, and is reported to have posted $7 million in revenue in 2003, which, if true, is quite good going after just three years' trading.

What is possible is that Sarvega was becoming vulnerable to an acquisition, and Intel decided to act before anyone less favorable to Intel's interests made their move. There are two separate reasons why Intel would want to see Sarvega continue its efforts along Intel-friendly lines:

  • Keeping XML processing on general-purpose platforms. Even though Intel's Tarari investment means it won't lose out entirely if Cisco has its way and XML processing migrates to dedicated chips embedded in network devices, Intel would prefer to see some of that action remaining on general-purpose platforms, either in dedicated appliances like Sarvega's or on general-purpose servers. Now that Intel owns Sarvega's "patent-pending XML EventStream Operating System(tm) (XESOS(tm)), pronounced 'ZEE SOS',", it has the option of, for example, making some of that technology available to be licensed for use by infrastructure software vendors such as IBM, BEA, Microsoft or even SAP (or Linux/Apache ...). That could help keep the XML processing from moving out of the sphere of Intel's existing markets.

  • Optimizing Sarvega's XML processing software for Intel Architecture. As an independent company, Sarvega made a point of keeping its software independent of any specific processor platform. Now that it's become part of Intel it can get on with the job of honing its performance on Intel chips, and maybe Sarvega had got to the point anyway where it had realized it was going to have to choose a platform if was going to remain competitive with its hardware-assisted competitors.

Does Intel's acquisition of Sarvega signal that it's gearing up to launch a new family of XML appliances, aggressively going after the same AON opportunity as Cisco? I had a quick chat last night with Chris Hogg, Intel's UK marketing manager, and what he said suggested that's unlikely. "I don't see it as a resurgence of a network appliance business [at Intel]," he told me. "I think optimizing the Sarvega product on the Intel Architecture is the key."

We'll see if Intel takes that further and actually starts licensing the technology to others. It depends really how much demand for XML processing is going to start emerging in the mainstream market. Intel seems to be betting that it's going to be important and has made this move to make sure it's ready to react if it needs to. Naturally that's delighted the incumbent vendors almost as much as Cisco's market entry in June. "Large players are seeing application aware networking as a key strategic technology," said Peter Ashton, VP product management at Solace Systems, in an email exchange last night. But Solace is betting squarely on the XML remaining at the network layer. "The network already touches all the bits anyway, why not have it do XML processing as well? You can’t do things like content-based QoS or IP VPN interworking unless your device is part of the network." While that's true, I suspect the reality is that other aspects of XML processing will sometimes happen elsewhere, depending on the application. Intel's Sarvega buy is designed to put it into a position to fully exploit those opportunities.

posted by Phil Wainewright 12:27 PM (GMT) | comments | link

Assembling on-demand services to automate business, commerce, and the sharing of knowledge

read an RSS feed from this weblog



latest stories RSS source

Headline news RSS source


Copyright © 2002-2005, Procullux Media Ltd. All Rights Reserved.