We evaluated 56-Kbps modem performance in terms of data compression and throughput. We tested over clean lines as well as over lines with various impairments. Because 56-Kbps modems rely on two competing and incom
patible chip sets -- K56flex and X2 -- we assembled two test-beds tailored to each technology. Before testing, we configured each modem to use hardware-based compression and the maximum connection rate supported by the modem's driver.
During a test session, we recorded the time required to download three files, each about 300 KB, and used the result to determine the Kbps throughput rate. The three files we used support a maximum compressibility of 1-to-1, 3-to-1, and 5-to-1, respectively. A serial port supporting baud rates of up to 230 Kbps was used for external modems.
Simulating a typical central-office connection was the TAS Series II telephone network emulator and Model 240
loop emulator
. To exercise the specific modems, we used the Total Control system from U.S. Robotics for testing X2 products, and the Max 4000 from Ascend for testing K56flex products.
To measure throughput under impaired conditions, we used the eight line conditions recommended by TAS for 56-Kbp
s modem testing; these conditions are currently in draft status before the Telecommunications Industries Association (TIA). The impairments represent various combinations of five factors: analog and digital pad loss, robbed-bit signaling, transhybrid loss, and delay. In addition, all lines (including the clean line used for comparison) were subject to the quantization noise, which normally occurs because of the analog/digital conversion involved with 56-Kbps modems.
The overall score for 56-Kbps modems comes from a 70:20:10 weighted rating of performance, usability, and features, respectively. We judged a modem's performance based on raw throughput on both clean and impaired lines. The better a modem's ability to compress data and deliver it at high speed, the higher its score.
ISDN Face-Off
For ISDN testing, our server had a high-speed serial port. A Teletone ISDN Simulator connected the server to the client PC. A high-speed serial port was also supplied on the client side to accommodate
the modem being tested.
As with the 56-Kbps modems, the time required to download the three 300-KB test files was recorded to determine the Kbps for each connection. The three files used supported a maximum compressibility of 1-to-1, 3-to-1, and 5-to-1, respectively. Several runs were performed on each file type and used to produce an aggregate score.
Evaluations in this report represent the judgment of BYTE editors, based on tests conducted by NSTL, Inc., as documented in a recent issue of its monthly
PC Digest
photo_link (
42 Kbytes)
240 loop emulator from Telecom Analysis Systems. Illustration is based on the 3Com Impact IQ.
Test Results
