Cell Relay Archive

Cell Relay Retreat>List Archive>month:1995-Apr> msg00409



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Fore setbatchsize?

  • From: walthalr@oneida.et.byu.edu (Robert W. Walthall)
  • Date: Tue, 25 Apr 1995 19:59:12 -0500, 25 Apr 95 18:58:06
  • Distribution: world

			Tx----Switch----Rcvr

I'm trying to monitor cell headers while transmitting at the maximum rate using Fore APIs.  We have a ForeRunner ASX 200, HP 700 series (80MHz variety) workstations (WS), and Fore HPA 200 series adaptors.  The cell headers can be viewed using aal_null. But when using aal_null we are dropping cells right and left.  I've done some reading and it seems that while the Tx rate is set by the PCR specified in the QoS at setup time, the receiving end must the use Fore's API atm_setbatchsize so as to transfer the c
orrect number of cells (from the input buffer?) to the upper (ATM?) layer (before the next sonet frame arrives and destroys the data?).  In which case it would be necessary to know the maximum Tx rate of the workstation in order to calculate the number of cells that will arrive per Sonet frame.

I would welcome advice as to the best/easist way to make that measurement, or if its already been done, a pointer to where I can find those numbers.

Also could someone please verify my method.  Assuming the WS to be capable of transmitting at 149.760 Mb/s (which I doubt). 

270 columns * 9 bytes = 2430 bytes
2430 - 90 (overhead)= 2340 bytes
2340 * 8 = 18.72 kbits
18.72 kbits / (1/8kHz) = 149.760 Mb/s

53 bytes * 8 = 424 bits

18.72 kbits / 424 bits = 44.15 cells per Sonet frame

Thus I would reduce the PCR to send exactly 44 cells/frame, and set the batchsize to 44.

Thanks in advance
Robert Walthall
Email walthalr@et.byu.edu
http://ccc.cs.byu.edu/~walthalr/