Part 2 of a two-part series.
When you get past the public images of Cisco – marketing machine, acquisitions juggernaut, Wall Street bellwether – there is a technology company. And a pretty good one, most users and analysts agree.
Part 1: Cisco learning to roll with the punches
Founded as a maker of boxes that connect computers, Cisco has expanded its scope into almost every area of corporate network computing, from LANs and WANs, to security, data centers, storage-area networks, telecom, voice and wireless. And as it branches out, Cisco’s technology strategy is shifting from an emphasis on individual products to more of a holistic system approach.
At the same time, Cisco watchers say the firm faces challenges as it transitions to a more internal development model rather than acquiring technology and products elsewhere.
IOS everywhere
The cornerstone of Cisco’s technology is its IOS software – the operating system that runs most of the company’s routing and intelligent switching gear for corporations and service providers.
While the evolution and expansion of IOS’s role over the years has led to much of Cisco’s success, the company also might face a daunting technical challenge: It’s widely installed, yet it’s 15 years old.
“Yes, it’s a monolithic piece of code that is sensitive to changes and new additions,” says Joe McGarvey, senior analyst for carrier infrastructure at Current Analysis. “At the same time, it runs throughout its product line and is ingrained in thousands of environments. It’s sort of like the weather: There’s no point in complaining about it because you can’t do anything about it.”
Analysts and competitors bash IOS for being monolithic, unstable and a CPU resource hog – Juniper often says that IOS saps wire-speed performance from Cisco 12000 -series routers when additional software-based services are turned on. Meanwhile, corporate customers say the myriad features and capabilities constantly being added make it tricky to manage a large IOS network.
“In many ways, IOS is good in that it drives change in terms of what IP networks can do,” says Chris Lukas, CTO of emerging technologies at Hold Brothers, a New York stock trading firm. But, he adds, “IOS has also been a thorn in my side. The biggest problem with it is that new releases aren’t always a superset of old releases. Things are left out in new versions that were there before.”
Dick Emford, lead network analyst for home-goods manufacturer Newell Rubbermaid in Freeport, Ill., says, “IOS is a pretty all-encompassing technology. Cisco is rolling everything up into it. As they start to roll out new functions, it’s kind of a challenge to find a stable version of IOS that also supports all the other functions we need.
“We do regression testing over and above what Cisco recommends for its IOS releases, and we try to minimize the flavors we keep,” he says. “It’s hard because every IOS version that we might have to roll out might have five or six different new functions that we’ll need to test.”
| ||||||||||||||||||||
Newell Rubbermaid has more than 500 routers and switches deployed running IOS throughout 70 sites worldwide, and uses the software to help deploy new features such as voice over IP (VoIP ) and localized content caching throughout its distributed enterprise.
The addition of quality-of-service and VoIP support in IOS is something Emford would like Cisco to continue. As Newell Rubbermaid rolls out a nationwide VoIP network, the tighter voice and data are integrated into Cisco products, the better, he adds.
Overall, Cisco has been consistent about integrating IOS onto new products to support more features, such as adding IOS to its Aironet 802.11 access points for security and management purposes, adding IOS-based VoIP capabilities to routers, to complement its Architecture for Voice, Video and Integrated Data IP telephony technology, and weaving IOS into its storage switches acquired from Andiamo .
Platform expansion
As the IOS software has evolved to include new features, Cisco has taken the same strategy with its hardware by adding capabilities to enhance and extend the life of its existing enterprise product lines.
Much of Cisco’s enterprise WAN and LAN product strategy and development has involved add-ons to core platforms – such as the 1700- to 3700-series routers, Catalyst switches, PIX firewalls and VPN concentrators. Such upgrades introduce new features to a product’s core function (such as adding VPN capabilities to a firewall) or completely recast the device (such as content-caching modules with hard drives for WAN routers).
Much of Cisco’s strategy for enterprise hardware rests on a few key platforms, such as the Catalyst 6500 . The switch line was a $3.3 billion revenue stream for Cisco in fiscal year 2002 – around 40% of its switch product revenue – and has brought in a total of $11.4 billion over the life of the product.
“Cisco is trying to get enterprises to think of long-term platforms to which Cisco can continually add value,” says Mark Fabbi, vice president and research director for Gartner. “The Catalyst 6500 is the best example of its strategy of selling large chassis that never have to be swapped out. Cisco will make continuous upgrades with new features, but they never want it to open up to a platform discussion.”
Since its introduction in 2000, the box has evolved from a 32G bit/sec LAN backbone switch to a multipurpose device, capable of acting as a WAN router, an intrusion-detection platform, voice trunking, content switching, firewall and VPN functions as well as 10G Ethernet.
“You may not need all capabilities today, but down the road they’ll be handy to have,” says Zeus Kerravala, an analyst with The Yankee Group
Upgrades in switching capacity and density – such as copper Gigabit Ethernet in 2000 and 10G Ethernet in 2001 – have come in parallel with feature upgrades, as the box has gone from 32G to 256G bit/sec, and recently a 720G bit/sec backplane was announced. Cisco engineers say the platform will be able to support 40G bit/sec in the future as bandwidth demand evolves.
The Catalyst 6500 continues to evolve with new features, upgrades and overhauls every 18 months or so, says Luca Cafiero, senior vice president and general manager for Cisco’s switching, voice and storage technology group.
R&D vs. M&A
The consistency of features across Cisco’s multitude of enterprise platforms has helped it become dominant in most of the markets it plays in. But some observers say the company is more of a technological follower than a leader.
“You haven’t noticed any great new products from Cisco as of late,” says Frank Dzubeck, president and CEO of Communications Network Architects. “What you see are a great deal of enhancements. They don’t seem to be spending money on making fundamental breakthroughs in technology.”
Hold Brothers’ Lukas says, “It bothers me a little that the [Catalyst 6500], the PIX and other key platforms were the result of an acquisition [from Crescendo in 1993]. But the fact is that they are good products. Cisco didn’t get their reputation completely through marketing and hype. They earned it. When the PIX first came out, and some customers said it stunk, [Cisco] saw the numbers. Then they went out and made a product that didn’t suck. They didn’t try to convince people that 100 million packet/sec was really 400 million. They made the PIX do 400.”
Cisco says its research-and-development spending has outpaced its acquisition spending over the years.
“There has always been a lot more interest from the outside world in acquisitions . . . and very little in [research and development],” said Charlie Giancarlo, Cisco’s senior vice president and general manager of product development, speaking at a Cisco media event earlier this month on switching technology. “But R&D has always been the majority of spending Cisco has done in new innovation.” He added that Cisco’s average R&D spending of $3.3 billion over the last 3 years is more than twice the amount of any of its competitors’.
R&D at Cisco will focus on making networks more intelligent – or, easier and less expensive to manage, Cisco executives say. The idea of building a network as a unified, interoperable system instead of a mesh of loosely interconnected boxes was recently outlined by Mario Mazzola, Cisco’s senior vice president for development.
In terms of development strategies at Cisco, there “is a move away from point products and a move toward global systems and solutions, Mazzola said. The intention is to make it easier to integrate different applications, and to that extent, networks are increasing their level of intelligence.”
Carrying on
On the service provider side, the products and customers take on different shapes, but the intent to move the market toward end-to-end IP networks – that is, Cisco networks – is the same.
Cisco is investing $10 billion over five years in its service provider business, a sum that is close to the five-year R&D budgets of telecom-only giants Lucent and Nortel . Key areas for that investment are core, edge and metropolitan IP routers and switches, specifically in a range of interfaces for edge functionality such as Ethernet, traditional TDM, and Layer 2 data service migration to Layer 3 backbones.
Roland Acra, Cisco senior vice president and service provider CTO, cites recent improvements, including scalability, reliability, and high-availability enhancements such as nonstop forwarding, stateful switchover and Globally Resilient Internet Protocol capabilities.
Metropolitan optical and voice are areas of significant investment for Cisco. Over time, Cisco believes TDM-based SONET and synchronous digital hierarchy rings will take on more packet-oriented capabilities. Meanwhile, IP telephony – service provider hosting and management of VoIP networks for corporations – and greenfield public voice service infrastructures are the key drivers in that category, Acra says.
“You can see us team up with the cable industry, the [multiservice operator], to build broadband voice capabilities since these guys are clearly on the offensive in that domain,” Acra says.”
But Cisco’s challenge in furthering its IP voice vision is this: Why would a carrier want to go through that when it has had a reliable, revenue-generating TDM infrastructure in place for close to 100 years?
In metropolitan optical, Cisco’s ONS 15454 SONET transport system is a popular product, with more than 900 customers and 30,000 systems deployed worldwide. But long-haul optical is another story.
Observers speculate that Cisco will exit the long-haul optical market where its ONS 15800 DWDM platforms have languished – one of the hardest-hit markets during the three-year telecom slump.
A router company at the core
Cisco continues to add high-speed edge capabilities to its 12000 series core router and position its 7600 series as the platform upgrade for the 8-year-old and widely installed 7500 series.
In core routers, Cisco’s market share in the fourth quarter of 2002 slid by 7%, to 73%, Synergy says. Meanwhile, rival Juniper saw its core router revenue climb nearly 40%, accompanied by a significant take-back in market share from Cisco. Juniper raised its core router market share 7.2%, to 24.8%, according to Synergy.
Juniper’s gain might be attributable to the unveiling of its T640 core router last year, which boasts terabit-level scalability, support for 40G bit/sec per slot, and a five-to-seven-year product lifespan, among other features. Meanwhile, Cisco’s 12000 series router line is 4 years old.
After Cisco unveiled the 12400 line in 2001, Juniper began losing share to Cisco. Now it has gained some back as the industry awaits Cisco’s new core offering, which might come later this year. Acra says a major 12000-series upgrade occurs every two years.
“No doubt about it, the Cisco 12000 series is long in the tooth,” Current Analysis’ McGarvey says. “However, Cisco is really smart about upgrading technology at the right time.”
Previously: Cisco learning to roll with the punches.
