News
Analysis: Cisco Throws Down the Gauntlet With New CRS-1
6/10/2004 -- One obvious upshot of Ciscos CRS-1 announcement is that the networking giant has gone on the offensive against its scrappier competitors for the first time in years.
That marks a reversal of course for Cisco, which months ago seemed content merely to tread water as newcomers Avici Systems Inc., Caspian Network, and Chiaro Networks, among others, peddled innovative new routers designed to accommodate the needs of the multi-service infrastructure. To counter this assault, Cisco introduced...an upgrade for its 12000 Series routers that allowed customers to double their existing capacity. The companys strategy, analysts said, was to bide its time and forestall customer defections until it could deliver an upgrade possibly in the form of a new system.
Now that this new system has more or less arrived, its not so hard to understand Ciscos months long élan -- even in the face of withering competition from a rejuvenated Juniper Networks.
After all, notes Jeff Ogle, an analyst with consultancy Current Analysis, the CRS-1 fills a gap in Ciscos product line that could not adequately be addressed by means of any conceivable upgrade to its 12000 Series.
[I]t establishes a new level of performance for a core router not only in terms of raw throughput, but also in software resiliency, availability, and extended product life cycles, he writes. The new product metrics for the CRS-1, especially the extended product life cycle, will change the business model for carriers and allow them to consider providing new service offerings that previously were technically infeasible or cost-prohibitive.
Whats so revolutionary about Ciscos new CRS-1? For starters, says Ogle, it sports a so-called Intelligent ServiceFlex design, which can segment traffic and network operations on a per-service or per-customer basis. This approach opens up new possibilities for telecommunications carriers, he notes: This system design allows carriers to adapt to changing customer needs and accelerate service delivery through configuration changes to the CRS-1s two most prominent new features: The next-generation IOS XR operating environment and an unprecedented 40-Gbps ASIC, the Cisco Silicon Packet Processor (SPP), which was co-developed with IBM Corp.
With any first-generation product of the CRS-1s scope, customers are bound to have concerns about stability and reliability. In this respect, Ogle notes, the partner testimonials that Cisco trotted out during the CRS-1 launch event were reassuring. The demonstration with & MCI & provided credibility that the CRS-1s development is complete and Cisco just needs a bit more time to roll out the product for its targeted [general availability] date in July, he writes.
The timing of Ciscos announcement is also important. After all, Ogle notes, Cisco has steadily ceded share to rival Juniper networks over the last several years, such that Juniper now controls 25.8 percent of the market, compared with Ciscos still dominant 71.1 percent. (Cisco once controlled 90 percent of this market), according to Synergy Researchs Q4 2003 Router Equipment Share report. The CRS-1 which Ogle concedes is a revolutionary new product obliterates the long-time performance edge that Juniper has enjoyed, and makes a strong case that Cisco is once again the vendor of choice for routing equipment up to and including, Ogle says, the routing core.
There are concerns, however. For starters, owners of Ciscos existing GSR product lines wont necessarily derive any benefits by deploying CRS-1. The total departure of the CRS-1 from the existing GSR product [lines] architecture means current Cisco customers gain no operational benefits or reduction in CapEx by deploying this platform, he writes. This will not sit well with customers having a large installed base of gear, but is a necessary step to allow Cisco to move forward.
Elsewhere, Ogle notes, the CRS-1 may simply be too much router for some customers. This is clearly not the router for every network and service provider and the architectural debate about a large core versus a fully meshed net can begin anew, he concludes. -Stephen Swoyer
|