Dave,

Mbyte depends on the system of standards that you are using.
In SI, IEEE, IEC  standards:
1 megabyte is 1000*1000 bytes
1 mebibyte is 1024*1024 bytes

http://members.optus.net/alexey/prefBin.xhtml
http://physics.nist.gov/cuu/Units/binary.html
http://en.wikipedia.org/wiki/Megabyte
etc.

Regards,
Eugene


On 9/25/07, Dave Plonka <[EMAIL PROTECTED]> wrote:
> On Mon, Sep 24, 2007 at 04:38:04PM -0300, Caio Brentano wrote:
> > Please, someone answer this stupid doubt.
> >
> > Any report have 3 basic values: packets, octets and flows. If I want to know
> > the traffic between 2 IP address, I must analyse the OCTETS value, right?
>
> If you want to estimate a traffic rate, in bits per second, yes.
> Packets per second and flows per second are also interesting, especially
> for appliation such as detecting anomalies such as denial-of-service
> attacks that are often invisible if you only consider bits per second.
>
> > This is the "amount of data" in each flow collect, right?
>
> Yes, at the IP level.  So remember to account for layer-2 headers
> and such when considering link capacity.
>
> > What I have to calculate to get this data in Mbytes ?
>
> Apart from the octet bit that has already been followed-up...
>
> Remember that we measure bandwidth/utilization in bits/per second.
>
> and that while 1KByte = 1024 bytes, and 1MByte = 1024*1024 bytes,
> 1Kbit = 1000 bits, and 1Mbit = 1000*1000 bits.
>
> So, bits per second is calculated by totaling up byte values from
> flow records and multiplying by 8, then converting to a rate by
> dividing by the collection interval in seconds (ie. 300 = 5 mins).
> (We also multiple by the sample rate if packet sampling is used
> for the specific flow export implementation you're using.)
>
> To get the appropriate metrix prefix for the magnitude, I use a perl
> subroutine called "scale" based on Tobi Oetiker's code to convert to
> "M", "k", or whatever prefix is appropriate.  Search for "sub scale"
> in here:
>
>    http://net.doit.wisc.edu/~plonka/FlowScan/new/CampusIO.pm
>
> I'll admit some early versions of my code had a problem with using
> 1024 rather than 1000, as I incorrectly thought bits were measured
> like bytes (with power of 2 multipliers).
>
> Dave
>
> As an aside, "Networks: A Systems Approach" by Larry Peterson and Bruce
> Davie is a good book with side-bars about this and other such topics.
>
>
> --
> [EMAIL PROTECTED]  http://net.doit.wisc.edu/~plonka/  Madison, WI
> _______________________________________________
> Flow-tools mailing list
> [EMAIL PROTECTED]
> http://mailman.splintered.net/mailman/listinfo/flow-tools
>
_______________________________________________
Flow-tools mailing list
[EMAIL PROTECTED]
http://mailman.splintered.net/mailman/listinfo/flow-tools

Reply via email to