adlink-ist / opensplice-tools Goto Github PK

ash-4.1$ pwd
/home/asmith/Downloads/opensplice-tools-master
bash-4.1$ ls
common.c LICENSE porting.h simpleidlpp testtypeSacDcps.d tglib.c
common.d lsbuiltin.c pubsub.c testtypeDcps.h testtypeSacDcps.h tglib.h
common.h lsbuiltin.d pubsub.d testtype.h testtypeSplDcps.c
genreader.c makefile README.md testtype.idl testtypeSplDcps.d
idl2md porting.c ShapeType.idl testtypeSacDcps.c testtypeSplDcps.h

bash-4.1$ make
In file included from common.h:4:0,
from pubsub.c:31:
/proj/sw/3rdparty/PrismTech/Vortex_v2/Device/VortexOpenSplice/6.6.2p2/HDE/x86_64.linux/include/dcps/C/SAC/dds_dcps.h:21:20: fatal error: c_base.h: No such file or directory
#include "c_base.h"
^
compilation terminated.
make: *** [pubsub.o] Error 1
bash-4.1$

Hi, I played around with this nice pubsub tool, and I find it very useful to write tests. In one of my test experiments I stumbled on the following.

In my experiment, I want a publishing node that publishes N key values in a single burst as fast as possible, and then wait forever. To achieve that I used the writer option N:R*B. Here, N is the number of key values, R is the rate, B is the burst size. Evidently, in my use case B equals N, because I want all data to be published at once. The question that I have relates to R (which is the rate in bursts/second). What is a proper values for R in my use case?

My naive was approach was to use R=0. After all, if I publish a data set of N key values only once and then wait forever, you would think that rate converges to 0 after waiting indefinitely. However, it looks like that data is published continously when R=0. Ofcourse, I can get my test case to work by using a very small R (e.g., R=0.001). In this case the set will be published once, and if my test finishes within 1000 sec I'll be OK. But it is not actually the same as my original problem, because I'll have be sure that my test now finishes within 1000 sec.

My concrete question is:
Is there any reason why N:0 * N does publishes data continuously, instead of publishing it once and then wait forever?