re: Cisco Ships OC192Froggy has it right. Both cisco and juniper are counting in the same way. When compared using the same counting method, cisco's new router actually provides almost twice the data rate and accomodates almost twice the number of OC-192 ports. Sure, juniper has been shipping the M160 a lot longer but part of their success has been based on "best of breed" claims. Looking forward, doesn't this mean that cisco now has "best of breed"? (Yeah, I'm ignoring avici in this discussion, and I'm assuming "best of breed" means higher data rates and more ports). Supposedly, juniper is developing the M320 which doubles the number of OC-192 ports than in the M160. However, this product when released just plays catch up with cisco's current product that is shipping. How can juniper claim a 12 month lead over cisco?
re: Cisco Ships OC192Steve: Not sure whether this is the right forum for my question. Can anyone comment about "Kestrel Solutions" and "Mayan Networks"? Kestrel's product (Add/Drop Multiplexer) is making use of FDM technology. Currently the only competition is from Centerpoint broadband technologies. I would like to know how good is Kestrel's management team, their chances of success in this overcrowded MAN market. The technology is not yet proven but they claim that they are in field trial with 2-3 RBOC's. I am considering a job offer from Kestrel and Mayan.
re: Cisco Ships OC192Just making it clear that its a marketing number, that's all... And, yes, "everybody" does it.
Yes, there are twice the OC192's in one Cisco 12416 than in one Juniper M160. However, that's not the only number that carriers care about. The fact that Juniper can put 2 M160's in a standard telco 7ft rack, whereas one Cisco 12416 takes up the entire 7ft rack, puts them on even ground when it comes to port density.
The 12416 isn't going to bring back Cisco customers that have already made the switch to Juniper. It will probably only serve to increase retention rate going forward.
re: Cisco Ships OC192"...Yes, there are twice the OC192's in one Cisco 12416 than in one Juniper M160. However, that's not the only number that carriers care about. The fact that Juniper can put 2 M160's in a standard telco 7ft rack, whereas one Cisco 12416 takes up the entire 7ft rack, puts them on even ground when it comes to port density.
The 12416 isn't going to bring back Cisco customers that have already made the switch to Juniper. It will probably only serve to increase retention rate going forward...."
From a port density point of view maybe it's marketing, but from and engineering point of view, Cisco has created a 320Gb/s switch fabric which allows 20Gb/s throughput to each slot.
So using a worse case scenario and saying that the max that can be sent to the switch fabric is 10gb/s for 15 slots(eventhough the GRPs send data to the switch fabric), that would mean 150gb/s of the switch fabric capacity is used, and more than half the capacity of the fabric remains.
To me that means scalability as new interface cards are developed. If Cisco made only a 150Gb/s switch fabric, then they'd have to upgrade to a newer higher capacity switch fabric if they came out with a two port OC-192 card.
re: Cisco Ships OC192We've clarified a story to make it more accurate, following a call from AT&T. AT&T says it is a "trial customer," not yet a regular paying customer, for Cisco's new routers.
re: Cisco Ships OC192"...Do you really believe that Cisco's new switch fabric will support 20Gb/s full-duplex per slot and that's where the 320Gb/s number comes from?..."
No, you're right, it's 10Gb/s (uni-directional), and 20gb/s (bi-directional). If they wanted to provide 20 gb/s (bi-directional) per slot, they'd have to upgrade the switch fabric.
Point taken....
I guess the real question here is whether or not it's fair to call a full-duplex 10gb/s port a 20gb/s port.
re: Cisco Ships OC192Could someone tell me the current projected market opportunties for JNPR in 2001? And also the projected market opportunities for JNPR in 3-5 years?
Not sure whether this is the right forum for my question. Can anyone comment about "Kestrel Solutions" and "Mayan Networks"? Kestrel's product (Add/Drop Multiplexer) is making use of FDM technology. Currently the only competition is from Centerpoint broadband technologies. I would like to know how good is Kestrel's management team, their chances of success in this overcrowded MAN market. The technology is not yet proven but they claim that they are in field trial with 2-3 RBOC's. I am considering a job offer from Kestrel and Mayan.
Yes, there are twice the OC192's in one Cisco 12416 than in one Juniper M160. However, that's not the only number that carriers care about. The fact that Juniper can put 2 M160's in a standard telco 7ft rack, whereas one Cisco 12416 takes up the entire 7ft rack, puts them on even ground when it comes to port density.
The 12416 isn't going to bring back Cisco customers that have already made the switch to Juniper. It will probably only serve to increase retention rate going forward.
The 12416 isn't going to bring back Cisco customers that have already made the switch to Juniper. It will probably only serve to increase retention rate going forward...."
From a port density point of view maybe it's marketing, but from and engineering point of view, Cisco has created a 320Gb/s switch fabric which allows 20Gb/s throughput to each slot.
So using a worse case scenario and saying that the max that can be sent to the switch fabric is 10gb/s for 15 slots(eventhough the GRPs send data to the switch fabric), that would mean 150gb/s of the switch fabric capacity is used, and more than half the capacity of the fabric remains.
To me that means scalability as new interface cards are developed. If Cisco made only a 150Gb/s switch fabric, then they'd have to upgrade to a newer higher capacity switch fabric if they came out with a two port OC-192 card.
Maybe it's forward thinking on Cisco's part?....
go_csco!!
Do you really believe that Cisco's new switch fabric will support 20Gb/s full-duplex per slot and that's where the 320Gb/s number comes from?
Do you have technical knowledge of the switch fabric implementation that supports this claim? Or are you inferring this from Cisco's marketing?
Until Cisco delivers a single-slot, 2-port OC-192 module or a single-slot 8-port OC-48, I ain't buyin' that story.
"then they'd have to upgrade to a newer higher capacity switch fabric"
Dosn't this make the 3rd GSR12000 switch fabric to support higher capacity?
http://staging.lightreading.co...
No, you're right, it's 10Gb/s (uni-directional), and 20gb/s (bi-directional). If they wanted to provide 20 gb/s (bi-directional) per slot, they'd have to upgrade the switch fabric.
Point taken....
I guess the real question here is whether or not it's fair to call a full-duplex 10gb/s port a 20gb/s port.
From a switch fabric point of view, traffic is ingress <tx> from the port on the OC-192c line card to the fabric, and egress <rx> from the fabric to line card . This represents one 10gb/s stream "east to west".
The opposite is true for traffic going "west to east" from line card to </rx>. This means the fabric only sees 20gb/s max stream between a pair of OC-192 line cards.
The above example is based on point to point sessions, the fabric does have the ability to replicate multicast packets.
I'm not totally clear on the logic, but Cisco is one in a crowd that counts this way.
I guess at the end of the day, the fabric has the ability to process 10gb/s per slot uni-directional, and 20gb/s per slot bi-directional.
The debate will rage on as to whether or not this is REALLY 320gb/s or 160GB/s.
go_csco!!
</tx>
The OC192 is half duplex, that means 10Gbps in
and 10Gbps out. That is why you need 10Gbps * 2
per slot = 320Gbps in the GSR12416.
DWDM