Mailing List Archive

On Wed, Dec 24, 2003 at 12:23:15PM +0100, Marcel Lemmen wrote:
> Hi,
>
> I'm a lucky owner (*sarcasm*) of a Cisco and at this point I'm hitting the
> Cisco 7500 Geip+ max performance. As you might know you can't achieve more
> then 400 Mbit/s RX and 400Mbit/s TX since the bus has been divided.
>
> Now we are looking for alternatives so my question, what is Juniper's
> interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX and
> 500 for TX or can one really achieve (close to) 1Gb/s?

I can do 1 Gbps, but it's not really fair to compare the ancient 7500 to
a Juniper M series, a compareable Cisco would be a 7600 or 12000 series.

/Jesper

--
Jesper Skriver, jesper(at)skriver(dot)dk - CCIE #5456

One Unix to rule them all, One Resolver to find them,
One IP to bring them all and in the zone to bind them.

On Wed, Dec 24, 2003 at 12:41:45PM +0100, Jesper Skriver wrote:
> > Now we are looking for alternatives so my question, what is Juniper's
> > interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX and
> > 500 for TX or can one really achieve (close to) 1Gb/s?
>
> I can do 1 Gbps,

How, with only 800mbps full-duplex interconnect PIC-FPC?


Regards,
Daniel

Wednesday, December 24, 2003, 1:07:36 PM, you wrote:
> On Wed, Dec 24, 2003 at 12:41:45PM +0100, Jesper Skriver wrote:
>> > Now we are looking for alternatives so my question, what is Juniper's
>> > interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX and
>> > 500 for TX or can one really achieve (close to) 1Gb/s?
>>
>> I can do 1 Gbps,

> How, with only 800mbps full-duplex interconnect PIC-FPC?

the PIC-FPC Slot Bandwidth for FPC Type 1 is 1Gbit/sec.

Josef



> Regards,
> Daniel
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp

I think it's 3.2Gbps per FPC, divided among the active PICs... there are 4
PIC slots, but if you add up to 2.5 Gbps of interface speeds, it is still
non-blocking. 2.5 Gbps of full-duplex traffic grows to near 3.2Gbps of
PIC-FPC traffic, because of J-Cell expansion.


Rubens


----- Original Message -----
From: "Daniel Roesen" <dr@cluenet.de>
To: <juniper-nsp@puck.nether.net>
Cc: "Jesper Skriver" <jesper@skriver.dk>
Sent: Wednesday, December 24, 2003 10:07 AM
Subject: Re: [j-nsp] Juniper's Gigabit performance


> On Wed, Dec 24, 2003 at 12:41:45PM +0100, Jesper Skriver wrote:
> > > Now we are looking for alternatives so my question, what is Juniper's
> > > interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX
and
> > > 500 for TX or can one really achieve (close to) 1Gb/s?
> >
> > I can do 1 Gbps,
>
> How, with only 800mbps full-duplex interconnect PIC-FPC?
>
>
> Regards,
> Daniel
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>

> Now we are looking for alternatives so my question, what is Juniper's
> interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX and
> 500 for TX or can one really achieve (close to) 1Gb/s?

It can go a bit further than 1 Gbps... on tests we violated mininum
interframe gap specs and the router accepted all traffic without loss.

> Could you please post your Juniper-experience to the list?

A strong feeling that packets travel thru C75xx in slow motion...


Rubens

On Wed, Dec 24, 2003 at 01:15:09PM +0100, Josef Buchsteiner wrote:
> the PIC-FPC Slot Bandwidth for FPC Type 1 is 1Gbit/sec.

Hm ok, I had otherwise in mind. Perhaps I did the mistake to divide
the 3.2Gbps FPC bandwidth by 4 (which would result in 800Mbps per
PIC).

Is your 1Gbps figure net, or gross (including J-Cell overhead)?


Regards,
Daniel

Wednesday, December 24, 2003, 2:42:33 PM, you wrote:
> On Wed, Dec 24, 2003 at 01:15:09PM +0100, Josef Buchsteiner wrote:
>> the PIC-FPC Slot Bandwidth for FPC Type 1 is 1Gbit/sec.

> Hm ok, I had otherwise in mind. Perhaps I did the mistake to divide
> the 3.2Gbps FPC bandwidth by 4 (which would result in 800Mbps per
> PIC).

> Is your 1Gbps figure net, or gross (including J-Cell overhead)?

There is no J-Cell going from the PIC towards the FPC so this
is net. The packet stream will be cut into Cell at the I/O
Manager towards Memory and this is where you have the memory
bandwidth of 3.2Gbps excluding J-Cell overhead for one I/O ASIC
and the reason why you will not be able to run line rate on
4 * 1port GE-PIC on one FPC which is also documented AFAIK.

thanks
Josef



> Regards,
> Daniel
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp

Out of curiosity, were you able to have the juniper transmit >1gbps as
well on a gig-e port , or only receive?
If so, how?

On Dec 24, 2003, at 7:36 AM, Rubens Kuhl Jr. wrote:

>
>> Now we are looking for alternatives so my question, what is Juniper's
>> interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX
>> and
>> 500 for TX or can one really achieve (close to) 1Gb/s?
>
> It can go a bit further than 1 Gbps... on tests we violated mininum
> interframe gap specs and the router accepted all traffic without loss.
>
>> Could you please post your Juniper-experience to the list?
>
> A strong feeling that packets travel thru C75xx in slow motion...
>
>
> Rubens
>
>
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>
--Phil Rosenthal
ISPrime, Inc.

On Dec 24, 2003, at 9:10 AM, Josef Buchsteiner wrote:
>
> There is no J-Cell going from the PIC towards the FPC so this
> is net. The packet stream will be cut into Cell at the I/O
> Manager towards Memory and this is where you have the memory
> bandwidth of 3.2Gbps excluding J-Cell overhead for one I/O ASIC
> and the reason why you will not be able to run line rate on
> 4 * 1port GE-PIC on one FPC which is also documented AFAIK.
>
I think there are some big misconceptions people have. Does this mean
that if you have 4 gig-e's doing random packet len, 0bps in one
direction, and 1000mbps in the other, it would max out at an aggregate
of ~3.2gbps or ~2.5gbps of actual IP data?

--Phil Rosenthal
ISPrime, Inc.

Only receive, output traffic was split between two interfaces in some tests,
filtered or rate-limited in others.


Rubens


----- Original Message -----
From: "Phil Rosenthal" <pr@isprime.com>
To: "Rubens Kuhl Jr." <rubens@email.com>
Cc: <juniper-nsp@puck.nether.net>; "Marcel Lemmen" <marcel@support.net>
Sent: Wednesday, December 24, 2003 6:20 PM
Subject: Re: [j-nsp] Juniper's Gigabit performance


> Out of curiosity, were you able to have the juniper transmit >1gbps as
> well on a gig-e port , or only receive?
> If so, how?
>
> On Dec 24, 2003, at 7:36 AM, Rubens Kuhl Jr. wrote:
>
> >
> >> Now we are looking for alternatives so my question, what is Juniper's
> >> interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX
> >> and
> >> 500 for TX or can one really achieve (close to) 1Gb/s?
> >
> > It can go a bit further than 1 Gbps... on tests we violated mininum
> > interframe gap specs and the router accepted all traffic without loss.
> >
> >> Could you please post your Juniper-experience to the list?
> >
> > A strong feeling that packets travel thru C75xx in slow motion...
> >
> >
> > Rubens
> >
> >
> >
> > _______________________________________________
> > juniper-nsp mailing list juniper-nsp@puck.nether.net
> > http://puck.nether.net/mailman/listinfo/juniper-nsp
> >
> --Phil Rosenthal
> ISPrime, Inc.
>

On Wed, Dec 24, 2003 at 03:20:29PM -0500, Phil Rosenthal wrote:
> Out of curiosity, were you able to have the juniper transmit >1gbps as
> well on a gig-e port , or only receive?
> If so, how?

We've had GE ports doing line-rate during DoS attacks.

Infact we've had 2 GE ports doing the ae0 (aggregated ethernet)
doing line rate. We've since gone to 10ge in those locations, so I can't
remember the exact hardware configuration.

Those were some impressive tests of the hardware. Vendors
that want hardware/forwarding testing just need to request a DoS
attack ;-)

- Jared

> On Dec 24, 2003, at 7:36 AM, Rubens Kuhl Jr. wrote:
>
> >
> >>Now we are looking for alternatives so my question, what is Juniper's
> >>interpretation of Gigabit. Is this 1000Mbps divided, 500Mbps for RX
> >>and
> >>500 for TX or can one really achieve (close to) 1Gb/s?
> >
> >It can go a bit further than 1 Gbps... on tests we violated mininum
> >interframe gap specs and the router accepted all traffic without loss.
> >
> >>Could you please post your Juniper-experience to the list?
> >
> >A strong feeling that packets travel thru C75xx in slow motion...
> >
> >
> >Rubens
> >
> >
> >
> >_______________________________________________
> >juniper-nsp mailing list juniper-nsp@puck.nether.net
> >http://puck.nether.net/mailman/listinfo/juniper-nsp
> >
> --Phil Rosenthal
> ISPrime, Inc.
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp

--
Jared Mauch | pgp key available via finger from jared@puck.nether.net
clue++; | http://puck.nether.net/~jared/ My statements are only mine.

On Dec 24, 2003, at 5:01 PM, Jared Mauch wrote:

> On Wed, Dec 24, 2003 at 03:20:29PM -0500, Phil Rosenthal wrote:
>> Out of curiosity, were you able to have the juniper transmit >1gbps as
>> well on a gig-e port , or only receive?
>> If so, how?
>
> We've had GE ports doing line-rate during DoS attacks.
>
> Infact we've had 2 GE ports doing the ae0 (aggregated ethernet)
> doing line rate. We've since gone to 10ge in those locations, so I
> can't
> remember the exact hardware configuration.
>
> Those were some impressive tests of the hardware. Vendors
> that want hardware/forwarding testing just need to request a DoS
> attack ;-)
>
> - Jared
>

Right, you're talking about inbound at 1gbps. My question was >1gbps
with nonstandard ethernet settings of transmit, not receive. I've had
1000mbit in, and 950mbit out simultaneously on gig-e cards on a M20...
Never more than that though (on a single card).
>
--Phil Rosenthal
ISPrime, Inc.

Wednesday, December 24, 2003, 9:22:30 PM, you wrote:
> On Dec 24, 2003, at 9:10 AM, Josef Buchsteiner wrote:
>>
>> There is no J-Cell going from the PIC towards the FPC so this
>> is net. The packet stream will be cut into Cell at the I/O
>> Manager towards Memory and this is where you have the memory
>> bandwidth of 3.2Gbps excluding J-Cell overhead for one I/O ASIC
>> and the reason why you will not be able to run line rate on
>> 4 * 1port GE-PIC on one FPC which is also documented AFAIK.
>>
> I think there are some big misconceptions people have. Does this mean
> that if you have 4 gig-e's doing random packet len, 0bps in one
> direction, and 1000mbps in the other, it would max out at an aggregate
> of ~3.2gbps or ~2.5gbps of actual IP data?


let me explain more in detail to clear things up. The memory
bandwidth from the I/O Manager is 32bit @125Mhz. which gives
you a raw bandwidth of 4Gbps. Since each J-Cell has an overhead
of 16 byte we come to the 3.2Gbps of J-Cell payload. From here
on we only talk about cells per second. So we can have at max.
per FPC an aggregate throughput of 4Gbps/((64*8)+(16*8))=
6,25Mcps.

So the question is IP-Payload identical to J-Cell Payload. Not
all the time and it depends on the packet-size. Lets do en
example.

64byte Ethernet Frame is 1488095,2 pps @46IP. So if you have a
stream of 4 PIC's and each sending this amount we need 1488096
cells * 4 = 5952384 Cells per second in total. So that should
work. Lets do an example of 1500 byte. Here we need 24 Cells
and with 1500byte@IP we can send max 81274,382 pps. So we need
81275 * 24 = 1950600 cells per second for one Stream. If you
multiply this by four you would need 7802400 cells/per second
which is 20% more what we can handle. ( I've not mentioned for
simplicity that we add some more J-Cells without ip-payload
which contain address information only for larger packets ).
Now you can do the academic maths. for all packets sizes.


hope this clarifies the topic
Josef





> --Phil Rosenthal
> ISPrime, Inc.

> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp