maybe this is a little what your talking about? not sayin hes right just happen to pick up
David A. Jackson
? Does increasing network latency mean it's time to buy component stocks?
Recently, multiple investors have asked us about the "latency
argument." In this note, we argue the answer is "no" and explain our
view that this argument is flawed on several levels.
? First, latency is not an indicator of increasing utilization of optical equipment
SONET and DWDM gear are designed to deliver voice and data in
slots of traffic, and there is generally no latency at this level of the
network unlike the IP layer, we think. So increased latency may be a
lead indicator of demand for IP routers, but not optical systems and
components.
? Second, the linkage between rising traffic and carrier capex is unclear
Over the past 12 months, carriers have redeployed systems and
swapped line cards from under utilized to congested routes. While
such substitution is unsustainable in the long-term, we think, it also
can delay a capex recovery and create sales volatility in the near-term.
? Third, utilization fluctuates widely as carriers consolidate
Since many carriers lease far more traffic than they fill with customers,
when they enter bankruptcy, their customers migrate to alternative
carriers and large amounts of bandwidth are freed up. Given our
expectation of further consolidation in the services group, forecasting
capacity utilization is likely to be a moving goal post.
? Fourth, and most importantly, spending today is driven by expected ROI
In the post-bubble world, capex is driven by expected return on
investment, not traffic growth or page views, we think. We believe
carriers will spend more on networks not when latency rises, but when
ROI in data networks improves.
? We have a Cautious view on Wireline Networking
Challenging fundamentals, including weak demand, lower carrier
capex and soft enterprise spending, could led to another leg down in
2H02-2003. |
