re: Procket Plows OnYou might be right there.... But if the company is making money they deserve a share of it.
Email Procket, no response. Call Juniper for a bug in the code, it takes them now a month to fix and get you a test image. If i yell at cisco they listen.
re: Procket Plows OnYour way off, we have both Juniper and Cisco. Procket must have been too busy to answer this time, we will try again.
What is the point of calling Tony Li, BJ or JC?? Everyone from top to bottom in all these companies should be impowered to help... Sales shouldn't get escalated to Execs unless its a price war.
re: Procket Plows OnjCurious why are you emailing Procket and Juniper if you only use Cisco gear? Sure Cisco listens, they have thousands of employees. Of course you are speaking to a customer service intern that will merely listen and then pass you on to a new salesguy to sell you the stuff to fix your problem your original sales guy told you would never need. Try calling john chambers. You can call Tony Li or Bobby Johnson direct and talk with them at least.
re: Procket Plows OnGet al these guys in the lab.. focus on vpls, vpns, martini, kompella, qos, security, and performance and see who wins. ------- No. Focus on the basics. Things like VPLS, VPNs, martini and kompella should NOT even be on the list. If you try and push features like that in, all your doing is creating a superficial check-the-box test which is next to useless. Or if you WANT to test some of those things, test them in a meaningful way for a core router. That means don't put in VRF testing, edge termination types of funcitons or encapsulations that are more appropriate for a multi-service edge test.
I do remember Light reading having a multi service test a while back. But I don't remember Juniper bothering to participate. While i didn't think much of the test, that was the appropriate one for all the MSE stuff.
I'm also more interested in availability and upgrade features which procket has rather than looking at a me-too list of edge features.
And as far as Juniper goes, I hope this time they don't get to influence the results of the test. i.e. packet reordering isn't "ok" and routing capacities will be measured by forwarding this time rather than by how much swap space the system has. And Juniper doesn't get to insert special tests to test their features (like traffic filtering last time).
As far as Juniper's BTexact test, the T640 test was a really/really short test of packet forwarding. The test was so constrained that it only really answered one question about small packet throughput.
re: Procket Plows OnIs LR going to hold another core router test? We really really need one again. Listening to all on these boards makes me laugh. lack of features, Procket density and performance, junipers bugs, etc.,
The T640s numbers have been publicly displayed by BTexact. Where are prockets? The T640 supports 64 192s per rack "today" and sonet link-bonding. Procket has been working on their solution since winny-ass and egotistical tony li left in 98 and they have what 25% more ports? Please please please let there be a day that carriers need 96 192s per rack... I will be even richer than I am today with my remaining stock!!!
Get al these guys in the lab.. focus on vpls, vpns, martini, kompella, qos, security, and performance and see who wins. Its been since around 99 or 00 since the last test.. we are due! And yes I like Juniper.. hate cisco.. and am sick of procket overly-hyped crap. This industry needed a second source and you got a hell of a good one with some of the best s/w engineers on the planet... quit bitching and be thank ful. When is the last time you got 60 or 70% off your core routers? It is ridiculous what cisco and juniper are doing to get our business these days.
re: Procket Plows OnI completely agree with skeptic about T640's testing by BTexact. It did not test switch performance by any stretch of imagination. packets came in via a port, were sent through the same destination port, how real is that?
And the strangest thing is that the report did not show the results for 41B packets even though it was explicitly mentioned that it was tested. Now, now, is there a "biggie" notch at 41B? Come on, come on, wherever you are!!!!
re: Procket Plows On>As far as Juniper's BTexact test, the T640 test was a really/really short test of packet forwarding.
Not to mention that Juniper paid for the results of the test... What we really need is a completely non-biased testing suite geared toward core router features - skip all of the converged platform bs and test things like availability, convergence, in-service upgrades, and scalability using a real world test environment. I would also like to see MPLS-to-QoS functionality with the ability to handle large volumes of disparate traffic types.
re: Procket Plows OnMost implementations have 'holes' at certain packet sizes that cause performance drops. I'm sure that Juniper didn't want this fact blatantly called out in a report they paid ~$150k for!
For example Cisco has a hole at 40B to 576B... (just kidding CSCO folks)
It's this type of omission/manipulation that shows how useless these types of tests are - as long as the vendors can manipulate what's being tested or the final results how can you believe anything within the report?
re: Procket Plows OnIf Procket is teetering, I can only wonder about the likes of Chiaro and Caspian. 'Killer Apps' Video on-demand and VoIP probably won't come fast enough to save them, the big players will just wait them out and pick through their remains. Reality bites.
BTW, the next time someone uses the 'apples vs oranges' comparison, throw this at them. Enjoy.
Email Procket, no response.
Call Juniper for a bug in the code, it takes them now a month to fix and get you a test image. If i yell at cisco they listen.
Procket must have been too busy to answer this time, we will try again.
What is the point of calling Tony Li, BJ or JC?? Everyone from top to bottom in all these companies should be impowered to help... Sales shouldn't get escalated to Execs unless its a price war.
Sure Cisco listens, they have thousands of employees. Of course you are speaking to a customer service intern that will merely listen and then pass you on to a new salesguy to sell you the stuff to fix your problem your original sales guy told you would never need.
Try calling john chambers. You can call Tony Li or Bobby Johnson direct and talk with them at least.
-------
No. Focus on the basics. Things like VPLS,
VPNs, martini and kompella should NOT even
be on the list. If you try and push features
like that in, all your doing is creating a
superficial check-the-box test which is next
to useless. Or if you WANT to test some
of those things, test them in a meaningful
way for a core router. That means don't
put in VRF testing, edge termination types
of funcitons or encapsulations that are more
appropriate for a multi-service edge test.
I do remember Light reading having a multi
service test a while back. But I don't
remember Juniper bothering to participate.
While i didn't think much of the test, that
was the appropriate one for all the MSE stuff.
I'm also more interested in availability
and upgrade features which procket has rather
than looking at a me-too list of edge features.
And as far as Juniper goes, I hope this time
they don't get to influence the results of
the test. i.e. packet reordering isn't "ok"
and routing capacities will be measured by
forwarding this time rather than by how much
swap space the system has. And Juniper doesn't
get to insert special tests to test their
features (like traffic filtering last time).
As far as Juniper's BTexact test, the T640 test
was a really/really short test of packet
forwarding. The test was so constrained that
it only really answered one question about
small packet throughput.
The T640s numbers have been publicly displayed by BTexact. Where are prockets? The T640 supports 64 192s per rack "today" and sonet link-bonding. Procket has been working on their solution since winny-ass and egotistical tony li left in 98 and they have what 25% more ports? Please please please let there be a day that carriers need 96 192s per rack... I will be even richer than I am today with my remaining stock!!!
Get al these guys in the lab.. focus on vpls, vpns, martini, kompella, qos, security, and performance and see who wins. Its been since around 99 or 00 since the last test.. we are due! And yes I like Juniper.. hate cisco.. and am sick of procket overly-hyped crap. This industry needed a second source and you got a hell of a good one with some of the best s/w engineers on the planet... quit bitching and be thank ful. When is the last time you got 60 or 70% off your core routers? It is ridiculous what cisco and juniper are doing to get our business these days.
Pat
And the strangest thing is that the report did not show the results for 41B packets even though it was explicitly mentioned that it was tested. Now, now, is there a "biggie" notch at 41B?
Come on, come on, wherever you are!!!!
was a really/really short test of packet
forwarding.
Not to mention that Juniper paid for the results of the test... What we really need is a completely non-biased testing suite geared toward core router features - skip all of the converged platform bs and test things like availability, convergence, in-service upgrades, and scalability using a real world test environment. I would also like to see MPLS-to-QoS functionality with the ability to handle large volumes of disparate traffic types.
For example Cisco has a hole at 40B to 576B... (just kidding CSCO folks)
It's this type of omission/manipulation that shows how useless these types of tests are - as long as the vendors can manipulate what's being tested or the final results how can you believe anything within the report?
about the likes of Chiaro and Caspian.
'Killer Apps' Video on-demand and VoIP probably won't come fast enough to save them, the big players will just wait them out and pick through
their remains. Reality bites.
BTW, the next time someone uses the 'apples vs oranges' comparison, throw this at them. Enjoy.
http://www.inno-vet.com/articl...
-lotb