Re: Transient Javaspace (Outrigger) Performance

John McClain <[email protected]> Mon, 28 Jul 2008 15:22:14 -0400
Newsgroups gmane.comp.java.sun.jini
Message-ID <[email protected]>
I am not sure I have ever measured the inter-write/take time before,
but these times strike me as a bit long for Transient Outrigger. The
write times look a bit slow too.

How big are the entries? How many iterations did you let the space and
clients warm up for? I asume the space isn't doing anything else?

Asaf Lahav wrote:

Hi
Everybody,

On a
performance test I conducted I got rather surprising results.

It
appears as if it takes quite a large time span between the time an
entry is
actually written to the space and until a 'taker' gets the entry
instance.

In
order
to perform the required measurements, I setup both the writer and the
taker on
the same machine but in separate JVMs.

The
'writer' application writes 3 types of entries and the 'taker'
application takes
every entry which comes into the space.

The
'writer' application prints out "System.currentTimeMillis()" to the
console before and after writing the entry.

The
'taker' application prints out "System.currentTimeMillis()" to the
console after taking the entry.

*It
is important to state that both the 'writer' JVM and the 'taker' JVM
both have
the entries class definitions on their class path. So no class
downloading is
done.

*It
is also important to state that the measurements taken were not first
time
invocations… a "warm up" process was conducted before measuring

Entry3

Entry2

Entry1

Samples

1217259525222

1217259524480

1217259522708

*writer
(before) tickcount

1

1217259525232

1217259524501

1217259522708

*writer
(after)

tickcount

1217259525612

1217259524631

1217259523559

*taker

tickcount

390

151

851

*taker-writer

(time span)

380

130

851

*taker-writer

(time span)

1217260189897

1217260188876

1217260188065

*writer
(before)

Tickcount

2

1217260189917

1217260188936

1217260188105

*writer
(after)

Tickcount

1217260190078

1217260189216

1217260188295

taker

Tickcount

181

340

230

*taker-writer
before

(time span)

161

280

190

*taker-writer
after

(time span)

Is
that reasonable performance for the Outrigger java-spaces
implementation?

Thanks
in advance,

Asaf

--------------------------------------------------------------------------
Getting Started: http://www.jini.org/wiki/Category:Getting_Started
Community Web Site: http://jini.org
jini-users Archive:
http://archives.java.sun.com/archives/jini-users.html
Unsubscribing: email "signoff JINI-USERS" to [email protected]

--------------------------------------------------------------------------
Getting Started: http://www.jini.org/wiki/Category:Getting_Started
Community Web Site: http://jini.org
jini-users Archive: http://archives.java.sun.com/archives/jini-users.html
Unsubscribing: email "signoff JINI-USERS" to [email protected]