x
Optical/IP

40-Gig Router Test Results

Test Objective

Determine the maximum forwarding performance for mixed IPv4 and IPv6 traffic.

Test Setup

For detailed information on the test setup, please see the complete Methodology, page 20.

In 2004, the percentage of IPv6 traffic in the Japanese Internet backbone reached 1 percent of all traffic. A worst-case guess is that IPv6 traffic might reach no more than 15 percent of all Internet traffic until 2010, which is the figure we used in this test. IPv6 is slowly getting rid of its research-network-only image: Carriers now require it for all new backbone equipment.

In our test, the CRS-1 had to prove that it processes IPv6 forwarding purely in hardware, and that the higher overhead to analyze the header does not lead to any performance issues – specifically in a realistic IPv4 and IPv6 environment. We prepared the Agilent emulators as in the first test case; the only difference was to substitute 15 percent of all traffic with IPv6. (The Internet-like mix for IPv6 also used different packet sizes.)

Results

The CRS-1 clearly proved that it processes IPv6 completely in hardware. In our mixed scenario, the single-chassis system mastered a packet rate of 820 million pps at line rate. The packet rate is slightly smaller than in the pure IPv4 scenario because IPv6 packets are longer.

Surprisingly, the system showed 1.83 percent loss in the first test run. Cisco explained that packet flows arriving at the Modular Services Card are processed by a chipset with 188 parallel channels, each of which can deal with one packet at a time. Under normal conditions, this is much more than enough for wire-speed throughput, as we have seen in all the other tests. The forwarding decision for an IPv6 packet takes 20 percent more time than for an IPv4 packet – our latency tests confirmed this value (see below). The scheduler assigning packets to each of the 188 channels has been optimized to perform at wire-speed for most IPv4 and IPv6 combinations. Our 85/15 percent case triggered a corner case by chance.

Cisco optimized the micro-code and installed a patch on the system under test. We reran the test case and verified that the CRS-1 now performed at 100 percent throughput without packet loss. The patch remained installed for all further test cases, to make sure that it didn’t harm the performance in other areas.



Table 1: Mixed IPv4/IPv6 packet rates
Transmit Rate (packet/s total) Receive Rate (packet/s total) IPv6 Transmit Rate (packet/s) IPv6 Receive Rate (packet/s) Packet Loss per Second Receive Bandwidth L2
First test run 819,904,533 802,170,750 130,840,624 115,837,061 15,003,563 (1.83 %) 98.17 %
Second test run, after micro-code change 819,927,902 819,927,902 130,840,608 130,840,608 0 100 %




The latency tests were run at 98 percent load and showed that the average latency is around 20 percent higher than in a pure IPv4 environment, and the maximum latency in a multi-chassis environment is 25 percent higher than in a single-chassis system. However, it is important to note that all IPv4 and IPv6 packets were forwarded within at most 0.34 milliseconds of time – this was the absolute maximum latency, which is a great result for IPv6.



Because of parallel packet processing in the Modular Service Card, we were aware of the theoretical possibility of out-of-sequence packets. EANTC ran a separate test again to verify that all packets within each flow were delivered in the correct sequence. In this test case (and in all of the others, too) Cisco maintained the sequence for each and every flow under any conditions. Also, there were no stray (misrouted) frames or any other unexpected issues. The implementation needs to be credited for the outstandingly precise scheduling in an ultra-high-speed parallel computing environment.

Next Page: IPv4/IPv6 IMIX-Based Forwarding With Security & QOS

Previous Page
4 of 11
Next Page
brahmos 12/5/2012 | 1:02:08 AM
re: 40-Gig Router Test Results I am sure many people will pick holes in the tests
but as usual csco has proved its dangerous to underestimate it.
could Alca, Lu or NT of similar size have done it ?
does anyone other than jnpr have a hope of matching
this beast?
beowulf888 12/5/2012 | 1:02:08 AM
re: 40-Gig Router Test Results Heh, heh. I guess people are still absorbing the "shock and awe" of it all ;-).

Of course, the CRS-1 is still an HFR (sorry, please don't fine me, Michael Powell!). They're definitely not going to do volume sales with this puppy. I wonder if the real threat from Cisco to other vendors is their taking some of this technology down to their other product lines.

cheers,
--Beo
trzwuip 12/5/2012 | 1:01:59 AM
re: 40-Gig Router Test Results hmmm, my 1$ worth here

Juniper can't make a competitive router in my opinion, QOS and m/cast on T640 are not good, 2 x 20g fabric does not constitute the ability for a real 40g port.

Sure, if you want a core box with no QOS / MCast etc then jnpr is fine, start adding these features combined with ACL's etc and watch your traffic rates die die die.

Read the recent report about 50% packet loss on the ERX
Belzebutt 12/5/2012 | 1:01:45 AM
re: 40-Gig Router Test Results And what does the ERX have to do with the T640.
russ4br 12/5/2012 | 1:01:41 AM
re: 40-Gig Router Test Results Read the recent report about 50% packet loss on the ERX

The ERX BRAS (ex-Unisphere) has its unique architecture and OS (JUNOSe). It has nothing in common with the T640, but for the Juniper brand.

-russ
beowulf888 12/5/2012 | 1:01:31 AM
re: 40-Gig Router Test Results trzwuip:
Is JNPR still being dogged by that packet reordering issue? (-- discussed in previous LightReading threads) If so, they've been plagued with that problem for a couple of years now.

JNPR has been way ahead of CSCO in their support for IPv6. I gather that IPv6 forwarding in hardware is still over a year out on some of CSCO's mainline platforms. But the CRS-1 changes this.

This might just change the dynamic in the Asian markets where IPv6 support is very very important.

--Beo
ragho 12/5/2012 | 1:01:19 AM
re: 40-Gig Router Test Results Actually, that's more like 1 cent worth than 1 $, if you ask me.
<ol>
<li>The competitiveness of Juniper is determined by customers, Juniper has numerous to prove it.
</li><li>I'm yet to see any vendor deliver a good multicast story. Cisco is no better, period.
</li><li>If you have some real data to share with regarding poor QoS in Junos, I'd be very interested in seeing that.
</li><li>If Cisco really made a box with 40G fabric ports, I'm sure we'll know about it. A 40G card doesn't make 40G port; most 10G cards are made today with 4x 2.5/3.125Gbps links into the fabric. 6.25/12.5Gbps serdes are still sampling these days (except for RaSer, which I know Cisco isn't using).
</li><li>The E-series is a different architecture than the M- and T-series. Please don't compare apples to oranges, or even cream cheese.
</li></ol>
My probable guess is that you haven't even touched a Juniper box, but want to claim otherwise. The ACLs in the M/T series are enforced at line rate in hardware, so your point about lowered traffic rates is just plain B/S. Get some real data and some wheaties.
Honestly 12/5/2012 | 12:58:46 AM
re: 40-Gig Router Test Results Well,
lets see. Test was done at Cisco, Agilent was trying to get anyone to sign a PO for a CRS for Cisco at Supercom and Cisco now has a program CALLED "STRATEGIC DONATION". True, They are the industry Santa Claus and are sending CRS-1's to carriers for FREE. Hey its a lot easier for Volpi to say the CRS is at carriers. Call and ask maybe they will raffle one off on LR. This then brings up the question. What did Cisco serve for dinner during the test, and was the wine good. ? Maybe we can get Zagat to give Cisco a 30, Extraodanary spending on anything that creates wishful buzz. After the CRS has run in a real production network with real traffic call me.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE