Juniper, Cisco battle to provide superior networking for Web services

If you see in any given situation only what everybody else can see, you can be said to be so much a representative of your culture that you are a victim of it.

S.I. Hayakawa

Dear Vorticians,

After sending out last week’s missive on the growing rivalry between Juniper and Cisco in the enterprise IT space, I received a terrific analysis of the situation entitled “Cisco, Juniper and … Sun?” from my connections at Eze Castle Research, headed up by Doug Ashton.

Eze Castle says the media is fostering an inaccurate – or at least insufficient – perception about the nature of the Cisco-Juniper battle. It’s not a contest between the approaches of one-stop-shopping and best-of-breed products. In fact, those are obvious statements of position for the two companies. Cisco, with a broad product line, is vulnerable to best-of-breed products stealing market share, particularly in high growth segments of the market, so it has to sell the mantra of end-to-end integration. For its part, Juniper can’t match Cisco’s portfolio, so it must emphasize the quality of specific point solutions and attack in the high-growth areas.

But the real, long-term battle, according to Eze Castle, is centered on the integration of applications and networking – more specifically, providing outstanding network support for the coming onslaught of Web services-based apps. (Precisely the issue that Geoffrey Moore and I explored in depth at VORTEX 2004, I might humbly add.)

“The key to all of this is Web services,” writes Eze Castle. “Described by Microsoft as ‘the application layer of IP,’ Web services is where networks and applications converge. IBM, Sun, Microsoft and pretty much every software vendor is rushing towards more Web-service-oriented offerings, and Web services clearly changes IP networking by adding a new layer.

“There have been LAN products for years purporting to be ‘Level x’ switches, where ‘x’ is some number higher than the Layer 2 or 3 where Ethernet switching and IP routing now live. The problem is that IP doesn’t have higher layers, or didn’t until Web services came along. There has been application switching and load balancing for years, but until Web services came along there was no reliable way to link traffic to applications. Firewalls have worked for years to keep out intruders, but without Web services’ identity management features, who is an intruder? In short, Web services touches every piece of enterprise networking and, through the managed service drives of carriers, every service provider deal as well.

“Cisco has been positioning itself as an element in Web services solutions for some time, but primarily as a suitable network platform, rather than an active participant. They have also been advocating the idea of ‘virtualization’ of resources in a number of forums, and that seems to be the focus of their TopSpin acquisition. But their only recent publication relating to Web services was a white paper on high-availability Web services that relates to an IBM application. Despite the lack of specific Web-service-based products to date, there’s been a long-standing rumor that Cisco intends to launch something. At their analyst day event, Chambers commented that future networks would ‘switch messages,’ the term used in Web services, and a Cisco engineering VP commented that the network could offload Web services handling from servers.

“Juniper has also been nibbling around the edges of Web services. The Infranet Initiative uses Web services for business control exchanges, and Juniper’s use of the term ‘Infranet’ in connection with its recent application access control and security announcement (the ‘Enterprise Infranet’) seems to indicate Juniper will be adopting Web services in its enterprise products as well. However, like Cisco, it has yet to make a definitive statement on Web services or field a product that is specifically targeted at the area.

“We think this is the real collision point between the two companies, one that neither can solve by small product-based acquisitions. To get to a higher level of strategic engagement with customers and to link their products to the current IT cycle, Cisco needs to have Web services as a bridge between IT and networking. To hop to the next area of strategic importance, Juniper needs to hop to Web services. Coming from two different directions, the firms are indeed on a collision course in the very specific product area of Web services support on the network.”

The research company lists a number of product areas that become critical in this battle (I can’t share everything for free!) and predicts a slew of announcements or acquisitions by Cisco and Juniper related to those segments in 2005 and 2006.

Now, if you’re wondering why Sun’s name was in the title of the report, the answer is that Eze Castle believes Scott McNealy’s company has some key assets that these network players would find very valuable.

“One tempting source of Web services hardware and software is Sun. The company’s price is so low that its multiple almost begs for an acquisition…. Sun is also the leader in providing servers to the telecom industry, the leading candidate for selling the servers that make applications like VoIP or IPTV work…. We’re not saying one or the other of these companies will buy Sun or another major IT player, just that maybe they should think pretty hard about it. The world – of networking, at least – is changing.”

Amen. Thanks to Eze Castle for letting me share ideas from its report. If you want to find out more, you can find the company at http://www.ezecastleresearch.com. If you want to share your own ideas, you can find me at mailto:jgallant@vortex.net.

Bye for now.