thousands of request in one port per second

[Praki]

Greetings All,

i think this is the right group to post this question.

i am working on client server model. in this model the client is
sending request in thousands in number. for example per sec around
2000 to 3000 request is coming. so the server is listening in the udp
port. after some time the most of the request is dropped. how can
manage this problem. is the multi threading will help ? but it is UDP
port i dont know how handle this request in multi threading model. or
is there any way i can handle this problem ?

thanks in advance

Prakash

[CBFalconer]

Praki wrote:

>
i think this is the right group to post this question.

i am working on client server model. in this model the client is
sending request in thousands in number. for example per sec
around 2000 to 3000 request is coming. so the server is listening
in the udp port. after some time the most of the request is
dropped. how can manage this problem. is the multi threading will
help ? but it is UDP port i dont know how handle this request in
multi threading model. or is there any way i can handle this
problem ?

Wrong newsgroup. We handle standard C, as described in the C
standard. The language contains none of clients, servers,
threading, UDP, etc. You probably need a newsgroup that deals with
your particular system.

--
[mail]: Chuck F (cbfalconer at maineline dot net)
[page]: <http://cbfalconer.home .att.net>
Try the download section.

[Antoninus Twink]

On 9 Oct 2008 at 14:32, Praki wrote:

i think this is the right group to post this question.

It's a perfectly appropriate group. Please ignore CBF, who is a known
troll, and senile to boot.

i am working on client server model. in this model the client is
sending request in thousands in number. for example per sec around
2000 to 3000 request is coming. so the server is listening in the udp
port. after some time the most of the request is dropped. how can
manage this problem. is the multi threading will help ? but it is UDP
port i dont know how handle this request in multi threading model. or
is there any way i can handle this problem ?

If dropped packets is a serious problem, you shouldn't be using UDP.

You could try increasing the size of the receiving buffer, e.g.:

int bufsize = 1<<16;
setsockopt(fd, SOL_SOCKET, SO_RCVBUF, &bufsize, sizeof bufsize);

If that fails, you could maintain your own buffer: your main thread
could just transfer data from the socket buffer to your buffer, and then
a worker thread could process items from your buffer. But ultimately, if
data is coming in faster than you can process it, then any buffer will
eventually overflow...

[Keith Thompson]

Praki <vi************ ***@gmail.comwr ites:

Greetings All,

i think this is the right group to post this question.

[...]

I'm afraid it really isn't. You'll bet better information from a
larger pool of experts in a group that deals with your operating
system, most likely comp.unix.progr ammer if you're on Linux or some
other Unix-like system, or comp.os.ms-windows.program mer.win32 or one
of the microsoft.* groups if you're on MS Windows.

I'm sorry that you've run into Antoninus Twink, a troll who sometimes
pretends to be helpful. If he were really interested in helping you
solve your problem rather than disrupting this newsgroup, he probably
would have told you about the existence of other more appropriate
groups.

--
Keith Thompson (The_Other_Keit h) ks***@mib.org <http://www.ghoti.net/~kst>
Nokia
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

[Richard]

CBFalconer <cb********@yah oo.comwrites:

Praki wrote:

>>
i think this is the right group to post this question.

i am working on client server model. in this model the client is
sending request in thousands in number. for example per sec
around 2000 to 3000 request is coming. so the server is listening
in the udp port. after some time the most of the request is
dropped. how can manage this problem. is the multi threading will
help ? but it is UDP port i dont know how handle this request in
multi threading model. or is there any way i can handle this
problem ?

Wrong newsgroup. We handle standard C, as described in the C
standard. The language contains none of clients, servers,
threading, UDP, etc. You probably need a newsgroup that deals with
your particular system.

Don't be ridiculous. Many aspects of these things can be done in standard
C and this is a good place to ask C programmers of their thoughts and
experiences. The fact that you have never programmed on a real
professional C project does not mean the rest here have not.

[user923005]

On Oct 9, 7:32*am, Praki <visitprakashin [e-mail] ote:

Greetings All,

i think this is the right group to post this question.

Probably not. I suggest news:comp.soft-sys.ace if you don't mind C++
solutions.
Using ACE would allow a fast, portable and reliable approach.
Building clients and servers with ACE is as easy as falling off a log.

i am working on client server model. in this model the client is
sending request in thousands in number. for example per sec around
2000 to 3000 request is coming. so the server is listening in the udp
port. after some time the most of the request is dropped. how can
manage this problem. is the multi threading will help *? but it is UDP
port i dont know how handle this request in multi threading model. or
is there *any way i can handle this problem ?

UDP gives no guarantees about delivery. If you need to scale lots of
requests, I suggest looking into memcached.
The memcached tool set is also a little unreliable (because it is a
cache), but it will scale to any load you can imagine.
http://en.wikipedia.org/wiki/Memcached

[Ian Collins]

Praki wrote:

Greetings All,

i think this is the right group to post this question.

i am working on client server model. in this model the client is
sending request in thousands in number. for example per sec around
2000 to 3000 request is coming. so the server is listening in the udp
port. after some time the most of the request is dropped. how can
manage this problem. is the multi threading will help ? but it is UDP
port i dont know how handle this request in multi threading model. or
is there any way i can handle this problem ?

If your server can't keep up, you should profile the code and see where
the bottlenecks are. It you fix those and it still can't keep up, use
faster hardware. If that can't keep up, use more cores and threading.
If that can't keep up, use a more appropriate protocol.

Or save a lot of time and do the last first.

--
Ian Collins

[Ian Collins]

user923005 wrote:

>
UDP gives no guarantees about delivery. If you need to scale lots of
requests, I suggest looking into memcached.
The memcached tool set is also a little unreliable (because it is a
cache), but it will scale to any load you can imagine.
http://en.wikipedia.org/wiki/Memcached

While memcached is a good tool, if the load is continuous and the server
can't keep up, the cache will fill and packets will continue to be dropped.

--
Ian Collins

[Ian Collins]

Richard wrote:

>
Don't be ridiculous. Many aspects of these things can be done in standard
C and this is a good place to ask C programmers of their thoughts and
experiences.

OK, put up or shut up. What are your thoughts and experiences? Or are
you just here to snipe form the sidelines?

--
Ian Collins

[user923005]

On Oct 9, 5:46*pm, Ian Collins <[e-mail] mwrote:

user923005 wrote:

UDP gives no guarantees about delivery. *If you need to scale lots of
requests, I suggest looking into memcached.
The memcached tool set is also a little unreliable (because it is a
cache), but it will scale to any load you can imagine.
http://en.wikipedia.org/wiki/Memcached

While memcached is a good tool, if the load is continuous and the server
can't keep up, the cache will fill and packets will continue to be dropped.

From the memcached FAQ:

"Finally, memcached itself is implemented as a non-blocking event-
based server. This is an architecture used to solve the C10K problem
and scale like crazy.

What's the big benefit for all this?
Carefully read the above entry (How does memcached work?). The big
benefit, when dealing with giant systems, is memcached's ability to
massively scale out. Since the client does one layer of hashing, it
becomes entirely trivial to add dozens of nodes to the cluster.
There's no interconnect to overload, or multicast protocol to implode.
It Just Works. Run out of memory? Add a few more nodes. Run out of
CPU? Add a few more nodes. Have some spare RAM here and there? Add
nodes!

It's incredibly easy to build on memcached's basic principles to
implement many different kinds of caching architectures. Hopefully
detailed elsewhere in the FAQ."