[j-nsp] LX SFP Question

Richard A Steenbergen ras at e-gerbil.net
Mon Sep 28 02:02:16 EDT 2009


On Sun, Sep 27, 2009 at 08:27:50PM -0700, joe mcguckin wrote:
> I recently saw an interesting article that claimed that Cisco consumed  
> 60-65% of the SFP market and thus was able to
> set pricing. Supposedly, Cisco pays about $25 for an SX SFP. Optical  
> modules represent one of the (if not the highest) margin product
> in their catalog.
> 
> No wonder they want to vendor lock the optics...

Actually, I'm pretty sure its way less than $25 for an SX SFP these
days. Of course the COGS on all router hardware is pennies compared to
the price charged (see Juniper's $20k tunnel pic with under $50 worth of
parts on it), the theory being that all the real money goes into R&D
(and taking the customers out for drinks :P) then the actual product
itself is dirt cheap to produce. The problem with the optics market is
the router vendors aren't even adding value to the product, they're just
reselling it for massive markup and rigging their routers to prevent any
competition. This would be like Ford voiding the warranty on your car
(or building a car that wouldn't even run) if you didn't buy tires your
tires from them, then marking the tires up 10000% since they have no
competition. Frankly I'm surprised someone hasn't stepped in to make
this illegal yet, other the ease of coming up with technobabble to
justify it to gullible lawmakers.

Another number I heard was that optic resale accounted for ~25% of
Cisco's yearly profit. With money like that, and the optics vendors
controlled by a clear monopsony, it's no wonder they're willing to
resort to vendor locking to keep people from realizing they could be
saving billions of dollars buying direct from the manufactures. At least
Juniper is one of the few companies to not engage in this downright
unethical practice, they deserve a lot of respect for that.

-- 
Richard A Steenbergen <ras at e-gerbil.net>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)


More information about the juniper-nsp mailing list