How to reliably measure the network bandwidth used by a process
Asked Answered
S

5

6

I have developed an application and I want to measure how much network bandwidth it consumes in some typical test cases.

I found a few applications like nethog etc. however I am not sure how accurate its reports are!

I would like some way to very accurately measure the same as the results need to go into a report for a conference.

I'm willing to write customized solutions for the same if someone guides me how to!

I want something where I can run the monitoring program and my target application to get network usage statistics - cumulative bytes sent/rcvd. .. maximum usage and average usage etc

Sacculate answered 24/2, 2012 at 18:58 Comment(3)
memory bus bandwidth? network bandwidth? what bandwidth?Decay
what makes you think nethogs is not accurate??Excreta
@MK I dont really know .. I am asking if it is reliable and not just a tool that gives a rough idea about how much each process is consumingSacculate
A
4

Can the application be isolated on a single machine?
Does anything else have to run on the system?

If a system can be dedicated this way, periodically grab the last line from /proc/net/netstat and subtract the corresponding values of InOctets and OutOctets.

This system, Fedora 15, shows this after 23 days of uptime:

TcpExt: SyncookiesSent SyncookiesRecv SyncookiesFailed EmbryonicRsts PruneCalled RcvPruned OfoPruned OutOfWindowIcmps LockDroppedIcmps ArpFilter TW TWRecycled TWKilled PAWSPassive PAWSActive PAWSEstab DelayedACKs DelayedACKLocked DelayedACKLost ListenOverflows ListenDrops TCPPrequeued TCPDirectCopyFromBacklog TCPDirectCopyFromPrequeue TCPPrequeueDropped TCPHPHits TCPHPHitsToUser TCPPureAcks TCPHPAcks TCPRenoRecovery TCPSackRecovery TCPSACKReneging TCPFACKReorder TCPSACKReorder TCPRenoReorder TCPTSReorder TCPFullUndo TCPPartialUndo TCPDSACKUndo TCPLossUndo TCPLoss TCPLostRetransmit TCPRenoFailures TCPSackFailures TCPLossFailures TCPFastRetrans TCPForwardRetrans TCPSlowStartRetrans TCPTimeouts TCPRenoRecoveryFail TCPSackRecoveryFail TCPSchedulerFailed TCPRcvCollapsed TCPDSACKOldSent TCPDSACKOfoSent TCPDSACKRecv TCPDSACKOfoRecv TCPAbortOnSyn TCPAbortOnData TCPAbortOnClose TCPAbortOnMemory TCPAbortOnTimeout TCPAbortOnLinger TCPAbortFailed TCPMemoryPressures TCPSACKDiscard TCPDSACKIgnoredOld TCPDSACKIgnoredNoUndo TCPSpuriousRTOs TCPMD6NotFound TCPMD5Unexpected TCPSackShifted TCPSackMerged TCPSackShiftFallback TCPBacklogDrop TCPMinTTLDrop TCPDeferAcceptDrop IPReversePathFilter TCPTimeWaitOverflow TCPReqQFullDoCookies TCPReqQFullDrop
TcpExt: 0 0 0 0 0 0 0 0 10 0 67116 0 0 0 0 8 117271 53 18860 0 0 102295 23352211 87967244 0 16861098 118195 893786 881659 0 29 10 0 0 0 9 10 16 12 2321 21 0 1 156 39 940 13 921 8015 0 1 2 0 18461 22 941 0 0 2974 15422 0 709 0 0 0 1 8 119 3 0 0 0 0 25231 0 0 0 4 0 0 0
IpExt: InNoRoutes InTruncatedPkts InMcastPkts OutMcastPkts InBcastPkts OutBcastPkts InOctets OutOctets InMcastOctets OutMcastOctets InBcastOctets OutBcastOctets
IpExt: 0 0 25308 48 725 1 24434248973 4218365129 2181277 13241 365505 65

Of course, that format is unfriendly for here, but fairly nice for scripting languages to deal with. You can see the depth and variety of information! The last line shows that this system has read 24,434,248,973 bytes and written 4,218,365,129. (It is on day nine of scraping a large website.)

Arlynearlynne answered 24/2, 2012 at 19:29 Comment(3)
Unfortunately another application using the network will have to run simultaneously on the same machine ! \Sacculate
@AnkurVj: Can the application be configured to do net i/o to only one other system? That is, could a dedicated server be set up to use the same technique, but from the other end?Arlynearlynne
yes I might be able to do that .. I think that should give me what I need .. thanks 1Sacculate
P
1

Vnstat is the simple tool to check the the internet bandwidth usage,

Here is the command to install it

sudo apt-get install vnstat

and to run the vnstat you need to run the below command.

vnstat

Here (check Monitor Network Bandwidth Usage) are some more option for checking daily, weekly, monthly, and top 10 day's network bandwidth usage.

Hope will help everyone.

Propend answered 15/8, 2013 at 19:4 Comment(0)
A
0

Poking around some more, I see that procfs contains what appears to be per process net i/o stats.

[wally@lenovotower ~]$ cat /proc/32089/net/dev
Inter-|   Receive                                                |  Transmit
 face |bytes    packets errs drop fifo frame compressed multicast|bytes    packets errs drop fifo colls carrier compressed
    lo:  622834    6102    0    0    0     0          0         0   622834    6102    0    0    0     0       0          0
  eth0: 3290609241 20752766    0    0    0     0          0         0 161708339 16831446    0    0    0     0       0          0
virbr0:       0       0    0    0    0     0          0         0        0       0    0    0    0     0       0          0

If this is a long-running process, then this could be used to calculate bandwidth used.

--- edit --- Despite the path, as others have pointed out, these are the same for all processes, and therefore obviously not per-process network i/o statistics.

Arlynearlynne answered 24/2, 2012 at 20:6 Comment(2)
I checked and found that the i/o stats displayed are not per process ... it is showing the same stats for all processes ...Sacculate
@Sacculate That's correct, /proc/<pid>/net/dev contains the same information as /proc/net/dev.Rightist
R
0

The only way of getting per task network io statistics would be using some kind of taskstats interface (based on netlink). Unfortunately it accounts everything you can imagine BUT network information. I have done a small patch for accounting bytes on a socket write/read and two entries (for tx/rx) on taskstats to get this kind of info from my systems.

Includes:

Signed-off-by: Rafael David Tinoco <[email protected]>
diff --git a/include/linux/taskstats.h b/include/linux/taskstats.h
index 341dddb..b0c5990 100644
--- a/include/linux/taskstats.h
+++ b/include/linux/taskstats.h
@@ -163,6 +163,10 @@ struct taskstats {
    /* Delay waiting for memory reclaim */
    __u64   freepages_count;
    __u64   freepages_delay_total;
+
+   /* Per-task network I/O accounting */
+   __u64   read_net_bytes;         /* bytes of socket read I/O */
+   __u64   write_net_bytes;        /* bytes of socket write I/O */
 };

And source code:

Signed-off-by: Rafael David Tinoco <[email protected]>
diff --git a/include/linux/task_io_accounting.h b/include/linux/task_io_accounting.h
index bdf855c..bd45b92 100644
--- a/include/linux/task_io_accounting.h
+++ b/include/linux/task_io_accounting.h
@@ -41,5 +41,12 @@ struct task_io_accounting {
     * information loss in doing that.
     */
    u64 cancelled_write_bytes;
+
+   /* The number of bytes which this task has read from a socket */
+   u64 read_net_bytes;
+
+   /* The number of bytes which this task has written to a socket */
+   u64 write_net_bytes;
+
 #endif /* CONFIG_TASK_IO_ACCOUNTING */
 };
diff --git a/include/linux/task_io_accounting_ops.h b/include/linux/task_io_accounting_ops.h
index 4d090f9..ee8416f 100644
--- a/include/linux/task_io_accounting_ops.h
+++ b/include/linux/task_io_accounting_ops.h
@@ -12,6 +12,11 @@ static inline void task_io_account_read(size_t bytes)
    current->ioac.read_bytes += bytes;
 }

+static inline void task_io_account_read_net(size_t bytes)
+{
+   current->ioac.read_net_bytes += bytes;
+}
+
 /*
  * We approximate number of blocks, because we account bytes only.
  * A 'block' is 512 bytes
@@ -26,6 +31,11 @@ static inline void task_io_account_write(size_t bytes)
    current->ioac.write_bytes += bytes;
 }

+static inline void task_io_account_write_net(size_t bytes)
+{
+   current->ioac.write_net_bytes += bytes;
+}
+
 /*
  * We approximate number of blocks, because we account bytes only.
  * A 'block' is 512 bytes
@@ -59,6 +69,10 @@ static inline void task_io_account_read(size_t bytes)
 {
 }

+static inline void task_io_account_read_net(size_t bytes)
+{
+}
+
 static inline unsigned long task_io_get_inblock(const struct task_struct *p)
 {
    return 0;
@@ -68,6 +82,10 @@ static inline void task_io_account_write(size_t bytes)
 {
 }

+static inline void task_io_account_write_net(size_t bytes)
+{
+}
+
 static inline unsigned long task_io_get_oublock(const struct task_struct *p)
 {
    return 0;
diff --git a/include/linux/taskstats.h b/include/linux/taskstats.h
index 341dddb..b0c5990 100644
--- a/include/linux/taskstats.h
+++ b/include/linux/taskstats.h
@@ -163,6 +163,10 @@ struct taskstats {
    /* Delay waiting for memory reclaim */
    __u64   freepages_count;
    __u64   freepages_delay_total;
+
+   /* Per-task network I/O accounting */
+   __u64   read_net_bytes;         /* bytes of socket read I/O */
+   __u64   write_net_bytes;        /* bytes of socket write I/O */
 };


diff --git a/kernel/tsacct.c b/kernel/tsacct.c
index 00d59d0..b279e69 100644
--- a/kernel/tsacct.c
+++ b/kernel/tsacct.c
@@ -104,10 +104,14 @@ void xacct_add_tsk(struct taskstats *stats, struct task_struct *p)
    stats->read_bytes       = p->ioac.read_bytes;
    stats->write_bytes      = p->ioac.write_bytes;
    stats->cancelled_write_bytes = p->ioac.cancelled_write_bytes;
+   stats->read_net_bytes   = p->ioac.read_net_bytes;
+   stats->write_net_bytes  = p->ioac.write_net_bytes;
 #else
    stats->read_bytes       = 0;
    stats->write_bytes      = 0;
    stats->cancelled_write_bytes = 0;
+   stats->read_net_bytes   = 0;
+   stats->write_net_bytes  = 0;
 #endif
 }
 #undef KB
diff --git a/net/socket.c b/net/socket.c
index 769c386..dd7dbb6 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -87,6 +87,7 @@
 #include <linux/wireless.h>
 #include <linux/nsproxy.h>
 #include <linux/magic.h>
+#include <linux/task_io_accounting_ops.h>

 #include <asm/uaccess.h>
 #include <asm/unistd.h>
@@ -538,6 +539,7 @@ EXPORT_SYMBOL(sock_tx_timestamp);
 static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
                             struct msghdr *msg, size_t size)
 {
+   int ret;
    struct sock_iocb *si = kiocb_to_siocb(iocb);
    int err;

@@ -550,7 +552,12 @@ static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
    if (err)
            return err;

-   return sock->ops->sendmsg(iocb, sock, msg, size);
+   ret = sock->ops->sendmsg(iocb, sock, msg, size);
+
+   if (ret > 0)
+           task_io_account_write_net(ret);
+
+   return ret;
 }

 int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
@@ -666,6 +673,7 @@ EXPORT_SYMBOL_GPL(sock_recv_ts_and_drops);
 static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct socket *sock,
                                   struct msghdr *msg, size_t size, int flags)
 {
+   int ret = 0;
    struct sock_iocb *si = kiocb_to_siocb(iocb);

    si->sock = sock;
@@ -674,7 +682,12 @@ static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct socket *sock,
    si->size = size;
    si->flags = flags;

-   return sock->ops->recvmsg(iocb, sock, msg, size, flags);
+   ret = sock->ops->recvmsg(iocb, sock, msg, size, flags);
+
+   if (ret > 0)
+           task_io_account_read_net(ret);
+
+   return ret;
 }

 static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
Retiary answered 14/8, 2013 at 17:23 Comment(0)
W
-1

Cat /proc/net/dev This is the one of the way to find out bandwidth usage

gaddenna@gaddenna-Vostro-3546:~$ cat /proc/net/dev 
Inter-|   Receive                                                |  Transmit
 face |bytes    packets errs drop fifo frame compressed multicast|bytes    packets errs drop fifo colls carrier compressed
 wlan1: 9420966650 7703510    0    1    0     0          0         0 673178457 4296602    0    0    0     0       0          0
  eth2: 7961371946 6849173    0   10    0     0          0    167030 446826449 3289015    0    0    0     0       0          0
    lo: 48209054   473527     0    0    0     0          0         0 48209054  473527     0    0    0     0       0          0
Waverley answered 15/11, 2016 at 19:21 Comment(0)

© 2022 - 2024 — McMap. All rights reserved.