Video test results
The video test was designed to demonstrate video prioritization performance of the AP/routers. We selected the three gigabit-capable routers for this test: D-Link DIR-655, Linksys WRT350N and Netgear WNR854T. These routers are high end models recommended by their manufacturers for HD video streaming and all three of them feature Gigabit Ethernet LAN ports with sufficient bandwidth to support a multimedia server sourcing multiple video streams.
The D-Link DIR-655 AP/router employs a proprietary QoS algorithm called Wireless Internet Stream Handling (WISH). WISH is implemented at the MAC layer alongside WMM1 and helps enhance video prioritization even when WMM is not used.
Video and background traffic was sent from two PCs connected to the Ethernet side of the router under test (Figure 19).
Click here for Figure 19
Figure 19: Video test setup—downstream traffic sent to the client to emulate the video stream from a video server that would typically be on the wired Ethernet network. The second Chariot endpoint on the Ethernet sent downstream background traffic.
The traffic was sent in the downstream direction to emulate a typical video transmission scenario where video may be streamed from a central server, such as Microsoft Home Server, to TV sets or set-top boxes throughout the house.
The test traffic was as follows. An http MPEG-2 video stream (137 MB file) was repeatedly sent from one of the Ethernet PCs and background traffic (throughput.scr) was sent from the other. The combined traffic from these PCs exceeded the available bandwidth on the Gigabit Ethernet side and on the WLAN side of the AP/router, requiring the router to select what traffic gets priority.
Prioritizing video traffic over background traffic is a key function of a WLAN router since video quality directly depends on the available throughput (table 2). For example, a 1080 progressive scan MPEG-2 HDTV stream that is refreshed at 30 frames per second requires 20 Mbps of throughput for good quality of video.
Click here for Table 2
Table 2: Video bandwidth requirements for common video formats and displays.
The throughput on the video transfer was measured and reported separately from the throughput on the background transfer to determine whether video traffic was allocated a larger percentage of the available bandwidth than background traffic.
The results reveal how effective D-Link DIR-655 is at prioritizing video traffic with respect to competing products from Linksys and Netgear (Figure 20).
Click here for Figure 20
Figure 20: Video throughput comparison of the AP/routers. The D-Link router provides sufficient throughput for an MPEG-2 HDTV stream even at 110 ft and through 5 walls. The competing Linksys and Netgear products would not support an MPEG-2 HDTV stream at this range.
The D-Link WISH algorithm is also shown to be very effective, allocating 100% more bandwidth to video traffic than to background traffic over a variety of physical layer conditions (Figure 21).
Click here for Figure 21
Figure 21: Video throughput of D-Link DIR-655 AP/router as a function of range. As total available throughput decreases with distance, video traffic maintains proportionally more bandwidth than background traffic. Even at 110 ft there is still sufficient video throughput for an MPEG2 HDTV stream.
The higher average throughput of video traffic can be seen on the time plots below (Figures 22-24).
Click here for Figure 22
Figure 22: D-Link DIR-655 AP/router at 6 ft.
Click here for Figure 23
Figure 23: D-Link DIR-655 AP/router at 40 ft.
Click here for Figure 24
Figure 24: D-Link DIR-655 AP/router at 110 ft—still sufficient throughput for an MPEG-2 HDTV stream.
Linksys WRT350N and Netgear WNR854T AP/routers offer no specialized prioritization for the video traffic and as a result the video throughput deteriorates proportionally to background traffic over distance as total available throughput diminishes (Figures 25-26).
Click here for Figure 25
Figure 25: Performance of the Linksys WRT350N AP/router at 110 ft " video throughput is at the same average level as the throughput of background traffic and video bandwidth is insufficient for transporting an MPEG-2 HDTV stream.
Click here for Figure 26
Figure 26: Netgear WNR854T AP/router at 110 ft. Netgear does not prioritize video traffic and, thus, provides insufficient throughput for MPEG-2 HDTV at this range.
Router test
The router test setup consisted of 2 Chariot endpoints—one connected at the LAN Gigabit Ethernet ports of the router and the other connected at the WAN Ethernet port (Figure 27). The Chariot throughput.scr script was used to measure throughput between the two ports.
Click here for Figure 27
Figure 27: Router test configuratio— bidirectional traffic was sent between two PCs, one on the WLAN side of the AP/router, the other on the WAN side.
The measurement results are summarized in Table 3.
Click here for Table 3
Table 3: Router test configuratio— bidirectional traffic was sent between two PCs, one on the WLAN side of the AP/router, the other on the WAN side.
The D-Link DIR-655 can route bi-directional TCP/IP packets at about 300 Mbps while performing Network Address Translation (NAT) and Stateful Packet Inspection (SPI). This high performance layer 3 routing is done by the Ubicom network processor integrated into the DIR-655.
Especially in Asia where deployment of fiber to the curb is becoming mainstream, 100Mbps+ throughput on the WAN is common and such high routing performance makes a big difference in the total application bandwidth available on the LAN.
Summary and Conclusions
This test, performed on draft 802.11n equipment, reveals more than 5x improvement in throughput with respect to the legacy 802.11a,b,g technology. The short range throughput was measured at nearly 140 Mbps for 3x3 devices, at around 110 Mbps for 2x3 devices and close to 100 Mbps for 2x2 devices.
Throughput measured in the office and in the home environments demonstrates that draft 802.11n equipment is competitive with the throughput of 100Base-T in a typical office or home setting.
Draft 11n equipment tested is well suited for video distribution in the home and can carry multiple video streams to WLAN-enabled displays and set-top boxes with comfortable coverage of a typical house. The WISH video prioritization algorithm implemented by the D-Link DIR-655 router further improves video quality and range by doubling the proportion of bandwidth allocated to video.
The impressive routing performance of the DIR-655 demonstrates that the Ubicom processor used in this product is number one in Gigabit-11n routing.
Among the products tested, the D-link DIR-655 with Atheros chipset came in first in its class of 3x3 MIMO and first in overall throughput, range and video performance. Linksys WRT350N was first in the 2x3 class and D-link DIR-625 was first in the 2x2 class.
Part 1 of this article is available at http://www.wirelessnetdesignline.com/howto/201800198.
Part 2 of this article is available at http://www.wirelessnetdesignline.com/howto/201800354.
A PDF version of the entire original report is available at http://www.octoscope.com/reports/octoscope_11n_test_19jun07.pdf.
About the author
Fanny Mlinarsky is an expert in wireless and RF data communications. She has served in key technology roles at Agilent, Azimuth and octoScope (www.octoscope.com), a Boston area consulting company.
----------
Footnotes:
1 WMM—Wireless Multi Media—the Wi-Fi Alliance QoS protocol that assigns different priorities to voice, video, background and best effort traffic.
|
