Well if we just look at LU and CSCO only the facts would seem to counter this position. LU
earned $2B on $32B in sales and holds a profit margin of 6.2%. Cisco generated nearly $2B
in earnings on only $11B in sales generating 3 times (18%) the profit.
Checking on this, doesnt sound right. LU has a lot of extraneous factors that could affect the raw numbers. Will get back to you on this.
These numbers are right off YAHOO.
The Pacific Rim is starting to dominate the low margin SOHO market that CSCO wants a
part of.
Really??? Who ???
Pick up a copy of warehouse.com's network catalog and you'll see a myriad of Pac Rim manufacturers practically giving this stuff away.
What products are in the catalog? What routers? What market? Are these products/markets that Cisco is targeting? You take the position that these far east manufacturers can kick Cisco's ass due to manufacturing inefficiencies..then you beat up Cisco for outsourcing to these same manufacturers? Pick a position Mizzou.
Hmmmm, $11B in invoices and no products.
I said MANUFACTURE, not design and sell, M-A-N-U-F-A-C-T-U-R-E. It can be very tricky to control costs when you outsource.
See above. Yes, Cisco does outsource much of its manufacturing. This reduces overhead thus increasing margins. A further increase to margins is the economies of scale derived by using these manufacturers. BTW: you might be interested to know that Cisco does continue to manufacture many products internally. But then you would have known that had you done your homework.
LU and NT will have some success in
the enterprise as Cisco has success in the SP's.
LU and NT will have more success in the enterprise than CSCO will in SP's. The size of the markets dictate this. Makes A LOT more sense for
CSCO to defend it's home turf vigorously than to give a little to get a little.
Well so far this doesn't appear to be the case. LU is virtaully unseen in the enterprise and NT still generates well over 90% of it's revenue from SP and 90% of that is circiut switches. Cisco nearly owns the cable market, well over 30% of the DSL DSLAM's and DSL deployments and continues to do well with ATM/IP products in the carrier space. I would argue that NT and LU are spending more time defending their turf than is Cisco. However, I would agree that since Cisco owns 80%+ of the enterprise that there is exposure.
Not only does Cisco support Tag...but they have the only ATM switch that is available with MPLS today.
Tag IS CSCO. It's proprietary and nobody's buying. They are WAY behind the game in ATM. MPLS isnt widely supported today so that isnt an
important issue yet.
Again, you have igonored the issue..the issue you brought up. Cisco has embraced both proprietary and standard based implementations. In the meantime LU and NT have nothing. Please explain to me how this puts them ahead? Again, Cisco demonstrates they are an innovator...not a follower...
We're going to rip out
the internet infrastructure and replace it with ATM, we're going to all install ATM NIC's on our
desktops and rip our our enterprise routers?
It is already happening. AT&T is replacing packet switching with ATM/IP. You know this OG, dont be so naive!
This is nothing different than AT&T has been doing for years. ATM core, packet (IP) edges. This has nothing to do with ATM to the desktop. In fact many SP's like UUNET is moving to all IP...some are moving to IP over SONET (POS).. then there is MPLS and Tag.. these technologies are all working toward delivering better services with the underlying assumption on IP edges.
Yes, ATM to the desktop is not
popular today, but gig ethernet is not living up to it's expectations either. As ASIC's make switches smarter and faster, the likelihood of the
demise of the traditional router is a foregone conclusion. This makes the header overhead of ATM less cumbersome and eventually practical for
desktop delivery.
Are you saying that ASIC's reduct the 5 byte overhead? Any idea what the latency is in a 53byte cell on a 56K line? Do the calculations my friend. Real time traffic wont fly. In fact point me to a spec that demonstrates ATM over anything less than T1 speeds. Point me to work that is actively going on that will allow ATM over low speed lines or to efforts underway to pull fiber all all the homes worldwide. I think 99.999% of those that know this industry and technolgy understand that IP edges are where it's at.
As for ATM in the campus - it makes no sense. The only benefit ATM has over IP today is QoS..and this differentiator is going away at RSVP, RTP, IP Presedence, DiffSev, etc. are implemented. Furthermore IP provides more flexibility allowing networks to be "appliation aware" - ATM can't do this...no one is even thinking about it. If there is delay in the campus it will cost far less to drop some fiber before it makes sense to install ATM. I can name 1000 customers installing IP on the campus for every 1 customer installing ATM. Is there any campus networks deploying pure ATM?...oh, perhaps a hospital somewhere...
I DONT expect this to happen within the next 5 years but I do expect it to happen long before it's pure IP end-to-end, if QoS can ever be supported in raw IP.
In 5 years the game will be over and I will be retired with homes in Tahoe, Phoenix, Monaco, and Santa Barbara.
OG
|
