Abstract

Cylink Corp. commissioned The Tolly Group to test the frames-per-second handling capacity of the Cylink AirLink (wireless) bridge and to demonstrate the bridge's ability to connect Ethernet LANs at extended distances and compare its performance to that of the Solectek AIRLAN/Bridge Plus. Both solutions use proprietary technology for long distance wireless transmission. 

In addition to performing baseline tests with the bridges co-located in The Tolly Group's lab, engineers performed testing at two set of other sites separated by distances of 2 and 12 miles respectively. The Cylink AirLink Bridge was able to deliver unidirectional throughput of 477.70 Kbit/s of 64-byte frames in all tests. 

The same tests were run on a Solectek AIRLAN/Bridge Plus remote wireless bridge. Co-located, the Solectek bridge forwarded unidirectional traffic at 251.90 Kbit/s. At 2 miles, that was reduced to 12.29 Kbit/s. As the Solectek wireless bridge is only rated at 3 miles, tests were not run as it was beyond the rated limit. 

In March 1995, The Tolly Group evaluated Cylink’s AirLink Bridge (ABE90) wireless Ethernet bridge—co-located and over field links of ~2.01 and ~12.22 miles in northern New Jersey—against Solectek’s AIRLAN/Bridge Plus. The focus was forwarding capacity with 64-byte frames and the ability to sustain throughput over distance.

Key results

• Cylink AirLink (ABE90):

  • Baseline (co-located): 1,866 fps total bidirectional (~955.40 kbit/s), i.e., ~933 fps per direction.

  • Unidirectional throughput: 477.70 kbit/s at baseline, 2.01 miles, and 12.22 miles; also maintained at 12.22 miles even when transmit power was reduced from 650 mW to 10 mW.

  • Bidirectional at 12.22 miles: sustained the same per-direction rate implied by the 477.70 kbit/s unidirectional figure.

• Solectek AIRLAN/Bridge Plus:

  • Baseline: 503.80 kbit/s total bidirectional (~251.90 kbit/s each way).

  • 2.01 miles: 12.29 kbit/s unidirectional in only one direction; reverse path failed; consistent 1 fps drop across offered loads.

  • 12.22 miles: no link (product rated for max ~3 miles).

• Filtering test (both products): While forwarding 100 fps, each processed 12,833 fps of local traffic—the generator limit.

Test setup & sites

• Two 10 Mbit/s Ethernet segments bridged over proprietary wireless links; analyzers measured frame rates.

• Field paths: 101 Hudson St., Jersey City ↔ 2 Journal Square (2.01 mi) and 101 Hudson St. ↔ Eagle Rock Reservation, Montclair (12.22 mi); thermal fade margin estimates were ~46.4 dB (2.01 mi) and ~30.6 dB (12.22 mi) for Cylink.

Notable specs (Cylink ABE90, vendor-supplied)

• Interfaces: AUI (10Base-5), 10Base-2, 10Base-T, EIA-530; bridging: Transparent, 802.1d STP; BOOTP/SNMP/TFTP with MIB-II and bridge MIB groups.

• Radio: 2.400–2.4835 GHz DSSS, BPSK, full-duplex TDM, 28 dBm max TX, -86 dBm @ 10^-6 BER.

• Antennas used in test: 24 dBi S-band semi-parabolic (Cylink) vs 13 dBi L-band Yagi (Solectek).

Bottom line: Cylink’s AirLink Bridge sustained ~477.70 kbit/s unidirectional throughput with 64-byte frames from lab conditions out to 12.22 miles, even at 10 mW TX power, while Solectek’s unit showed limited distance performance and directional instability beyond baseline.