Fwctl::AcctReport - Generates reports from C<fwctl dump-acct> command.


Fwctl documentation Contained in the Fwctl distribution.

Index


Code Index:

NAME

Top

Fwctl::AcctReport - Generates reports from fwctl dump-acct command.

SYNOPSIS

Top

    use Fwctl::AcctReport;

    my $report = new Fwctl::Report( options ... );

    my $r = $report->summary_report;

    foreach my $r ( @$src_alias_sum ) {
	print $r->{name}, " = ", $r->{bytes_sum}, "\n";
    }

DESCRIPTION

Top

The Fwctl::AcctReport(3) module can be used to generate packet and bandwith report from the data generated by the fwctl dump-acct command.

This module can be used as backend for two kinds of report. Summary which compiles the total number of packets and bytes received by each accounting chains over the report's period or histogram type report where you have the packets and bytes total for each sample (hour, day, etc) over the report's period.

CAVEATS

Top

The fwctl dump-acct commands dumps the packets and bytes total since the time of the last invocation. If the last invocation wasn't wholly comprised in the current sample, the total will be interpolated. This means that you should run the fwctl dump-acct command at least as frequently as the smallest sample period you want to use.

CREATING A NEW REPORT OBJECT

Top

    Ex. my $report = new Fwctl::AcctReport( start  => 'yesterday',
					    period => '1 day',
					    sample => '1d',
					    files  => [ 'log' ] );

PARAMETERS

The new method accepts the following parameter :

files

Specifies the file from which to read the dumped acctounting informations. It is an array of file handle or file names. If this parameter is not specified the records will be read from STDIN.

start

Sets the start of the report's period. If the Date::Manip(3) module is installed, you can use any format that this module can parse. If that module isn't installed you must use the following format YYYY-MM-DD HH:MM:SS or any meaningful subset of that format.

If this option is not used, the report will start with the first record.

end

Sets the end of the report's period. If the Date::Manip(3) module is installed, you can use any format that this module can parse. If that module isn't installed you must use the following format YYYY-MM-DD HH:MM:SS or any meaningful subset of that format.

If this option is not used, the report will end with the last record.

period

Sets the length of the report's period. This length is interpreted relative to the report's start. This option has priority over the end option.

If you have the Date::Manip module installed, you can use any format that this module can parse. If that module isn't available, you can use a subset of the following format X weeks X days X hours X mins X secs.

sample

Sets the length the sample's for the report. The sample_report will returns results aggregated over that length. For example, if you use 1d as sample's length, you will get the stats for every day from the beginning of the report until the end.

If you have the Date::Manip module installed, you can use any format that this module can parse. If that module isn't available, you can use a subset of the following format X weeks X days X hours X mins X secs.

names

Restrict records to those whose name matches names.

You can use this parameter multiple times to specify multiple possibility. The record will be included if it matches any of those.

METHODS

Top

start()

Top

Return the start of the report in seconds since epoch.

end()

Top

Returns the end of the report in seconds since epoch.

period()

Top

Returns the length of the report's period ( $report->end() - $report->start() )

sample()

Top

Returns the length of the sample period.

records()

Top

Returns an array reference to all the records read and which makes the report's samples.

RECORD FIELDS

Each record is an array ref. You can accessed the individual fields of the record by using the following constants. (Those can be imported by using the :fields import tag.)

TIME

The epoch start time of the sample.

NAME

The name of the accounting chain.

PACKETS

The number of packets received.

BYTES

The total number of bytes.

REPORTS

Top

The following report generation methods are available :

summary_report()

    my $r = $report->summary_report();




Generates a report that shows the compiled statistics for every accounting chain.

The resulting report is an array ref of hash reference. Each report record's has the following fields.

name

The name of the chains.

packets_sum

The total number of packets over the report's period.

This number is a Math::BigInt(3) object.

bytes_sum

The total number of bytes received over the report's period.

This number is a Math::BigInt(3) object.

packets_avg

The average number of packets received by that chain by sample.

This number is a Math::BigInt(3) object.

bytes_avg

The average number of packets bytes by that chain by sample.

This number is a Math::BigInt(3) object.

sample_report()

    my $r = $report->sample_report();




Generates a report that compiles the stats for each sample for each chains.

The resulting report is an hash reference. Each key is a chain and each value is an array ref to the chain compiled statistics. Each element of the array is an hash reference comprises of the following elements :

start

The start time in seconds since epoch for that sample.

packets

The total number of packets received in that sample.

This number is a Math::BigInt(3) object.

bytes

The total number of bytes received in that sample.

This number is a Math::BigInt(3) object.

AUTHOR

Top

Francis J. Lacoste <francis.lacoste@iNsu.COM>

COPYRIGHT

Top

SEE ALSO

Top

Fwctl(3) Fwctl::RuleSet(3) fwctl(8) fwctllog(8) fwctacctreport(8) Fwctl::Report(3) Date::Manip(3).


Fwctl documentation Contained in the Fwctl distribution.
#
#    Report.pm - Module that compiles accounting reports from the data dumped
#		 by the fwctl dump-acct command.
#
#    This file is part of Fwctl.
#
#    Author: Francis J. Lacoste <francis.lacoste@iNsu.COM>
#
#    Copyright (C) 2000 iNsu Innovations Inc.
#
#    This program is free software; you can redistribute it and/or modify
#    it under the terms of the GNU General Public License as published by
#    the Free Software Foundation; either version 2 of the License, or
#    (at your option) any later version.
#
#    This program is free software; you can redistribute it and/or modify
#    it under the terms same terms as perl itself.
#
package Fwctl::AcctReport;

use strict;

use vars qw( $VERSION @EXPORT @EXPORT_OK %EXPORT_TAGS @ISA );

use Math::BigInt;
use Symbol;
use Fwctl::Report;
use Exporter;

BEGIN {
    ($VERSION) = '$Revision: 1.4 $' =~ /(Revision: ([\d.]+))/;
    @ISA = qw( Exporter );

    @EXPORT = ();

    @EXPORT_OK = ();

    %EXPORT_TAGS = ( fields => [ qw(  TIME PACKETS BYTES NAME ) ],
		   );
}

BEGIN {
    # Create the necessary constant
    my $i = 0;
    for my $f ( @{$EXPORT_TAGS{fields}} ) {
	eval "use constant $f => $i;";
	$i++;
    }

    Exporter::export_ok_tags( 'fields' );
};

sub read_records {
    my $self = shift;

    # Read in the data
    my $records = [];
    push @{$self->{opts}{files}}, \*STDIN unless @{$self->{opts}{files}};
  FILE:
    foreach my $file ( @{$self->{opts}{files}} ) {
	my $fh;
	if ( ref $file ) {
	    $fh = $file;
	} elsif ( $file eq "-" ) {
	    $fh = \*STDIN;
	} else {
	    $fh = gensym;
	    open $fh, $file
	      or do { warn "can't open file $file\n"; next FILE };
	}

	# We subtract gives ourselves more records since the entries are
	# the amount between the current record timestamp and the one before
	# it.
	my $start = $self->{start};
	$start -=  2 * $self->{sample} if defined $start;

	my $end = $self->{end};
	$end += 2 * $self->{sample} if defined $end;

      LINE:
	while  ( my $line = <$fh>) {
	    chomp $line;
	    my @fields = (split /\s+/, $line)[0,2,3,4];

	    if ( ! defined $self->{start}) {
		$start = $self->{start} = $fields[TIME];
		$start -= 2 * $self->{sample};
		if ( defined $self->{period} ) {
		    $end = $self->{end} = $self->{start} + $self->{period};
		    $end += 2 * $self->{sample};
		}
	    }
	    # Skip fields outside the period
	    next unless $start <= $fields[TIME] && $end >= $fields[TIME];

	    if ( $self->{opts}{names} ) {
		foreach my $chain ( @{ $self->{opts}{names} } ) {
		    if ( $chain eq $fields[NAME] ) {
			push @$records, \@fields;
			next LINE;
		    }
		}
	    } else {
		push @$records, \@fields;
	    }
	}
    }

    # Records are sorted by time.
    $self->{records} = [ sort { $a->[TIME] <=> $b->[TIME] } @$records ];
}

sub new {
    my $proto = shift;
    my $class = ref $proto || $proto;

    my $self = { opts	    => { @_ },
		 records    => undef,
		 start	    => undef,
		 end	    => undef,
		 period	    => undef,
		 sample	    => 3600, # One hour sample
		 sample_report  => undef,
		 summary_report => undef,
	       };

    # Determine start and end of the report;
    if ( $self->{opts}{start} ) {
	$self->{start} = Fwctl::Report::parse_date( $self->{opts}{start} )
	  or die "invalid start date format: $self->{opts}{start}\n";
    }

    if ( $self->{opts}{period} ) {
	$self->{period} = Fwctl::Report::parse_period( $self->{opts}{period} )
	  or die "invalid period delta: $self->{opts}{period}\n";
	if (  $self->{start} ) {
	    $self->{end} = $self->{start} + $self->{period};
	}
    } elsif ( $self->{opts}{end} ) {
	$self->{end} = Fwctl::Report::parse_date( $self->{opts}{end} ) 
	  or die "invalid end date format: $self->{opts}{end}\n";
    } else {
	$self->{end} = time;
    }

    if ( $self->{opts}{sample} ) {
	$self->{sample} = Fwctl::Report::parse_period( $self->{opts}{sample} )
	  or die "invalid sample period: $self->{opts}{sample}\n";
    }
    bless $self, $class;

    $self->read_records;

    $self;
}

sub start {
    $_[0]->{start};
}

sub end {
    $_[0]->{end};
}

sub period {
    $_[0]->{end} - $_[0]->{period};
}

sub sample {
    $_[0]->{sample};
}

sub records {
    # Copy the records
    [ @{$_[0]->{records}} ];
}



sub summary_report {
    my ( $self ) = shift;

    return $self->{summary_report} if defined $self->{summary_report};

    my %chains = ();
    my $sample_report = $self->sample_report;
    while ( my ( $chain, $samples ) = each %$sample_report ) {
	my $stats =  {
		      name	      => $chain,
		      packets_sum   => new Math::BigInt( 0 ),
		      bytes_sum     => new Math::BigInt( 0 ),
		      packets_avg   => 0,
		      bytes_avg     => 0,
		     };

	foreach my $r ( @$samples ) {
	    $stats->{packets_sum} += $r->{packets};
	    $stats->{bytes_sum} += $r->{bytes};
	}
	$stats->{packets_avg} = $stats->{packets_sum} / @$samples;
	$stats->{bytes_avg}   = $stats->{bytes_sum} / @$samples;

	# Replace samples by stats
	$chains{$chain} = $stats;
    }

    # Returns an array ref sorted by name
    return $self->{summary_report} = [ sort { $a->{name} cmp $b->{name} } values %chains ];
}

sub sample_report {
    my ( $self ) = shift;

    # Return cache
    return $self->{sample_report} if defined $self->{sample_report};

    my $start  = $self->{start};
    my $end    = $self->{end};
    my $sample = $self->{sample};
    my $num_samples = int( ($end - $start ) / $sample );
    $num_samples++ unless $num_samples * $sample + $start > $end;

    my %chains = ();
    # Sort records by chain, since records are ordered by time,
    # all chain's record are also ordered.
    foreach my $r ( @{ $self->{records} } ) {
	push @{$chains{$r->[NAME]}}, $r;
    }

    # Foreach chains build an array containing the number of bytes and packets
    # by sample period.
    while ( my ( $chain, $records ) = each %chains ) {
	# Initialize the samples array for that chain
	my @samples = ( ) x $num_samples;
	for ( my $i = 0 ; $i < $num_samples; ++$i ) {
	    $samples[$i] = {
			    start   => $start + $i * $sample,
			    packets => new Math::BigInt(0),
			    bytes   => new Math::BigInt(0),
			   };
	}

	my ($last_s_idx, $last_time) = ( -1, 0 );
      RECORD:
	foreach my $r ( @$records ) {
	    # Find sample in which this record falls
	    my $s_idx = int ( ( $r->[TIME] - $start ) / $sample );

	    # If this record is before the beginning of the
	    # period, save time for normalization and go to
	    # next record
	    if ( $s_idx < 0 ) {
		$last_time = $r->[TIME];
		next RECORD;
	    }

	    # Comparing to the last record, is it completely
	    # in this period.
	    if ( $last_s_idx == $s_idx ) {
		# Add bytes and packet counters
		$samples[$s_idx]{bytes}   += $r->[BYTES];
		$samples[$s_idx]{packets} += $r->[PACKETS];
	    } else {
		# Distribute pro rata
		my $this_sample_ratio = ($r->[TIME] - ($start + $s_idx * $sample) ) / $sample;
		my $last_sample_ratio = 1 - $this_sample_ratio;
		if ( $s_idx >= $num_samples ) {
		    # This is the last record
		    $samples[$last_s_idx]{bytes}   += int( $r->[BYTES]   * $last_sample_ratio );
		    $samples[$last_s_idx]{packets} += int( $r->[PACKETS] * $last_sample_ratio );
		    last RECORD;
		} elsif ( $s_idx == 0 ) {
		    # This is the first record
		    $samples[$s_idx]{bytes}   += int( $r->[BYTES] * $this_sample_ratio );
		    $samples[$s_idx]{packets} += int( $r->[PACKETS] * $this_sample_ratio );
		} else {
		    $samples[$last_s_idx]{bytes}   += int( $r->[BYTES] * $last_sample_ratio );
		    $samples[$last_s_idx]{packets} += int( $r->[PACKETS] * $last_sample_ratio );
		    $samples[$s_idx]{bytes}   += int( $r->[BYTES] * $this_sample_ratio );
		    $samples[$s_idx]{packets} += int( $r->[PACKETS] * $this_sample_ratio );
		}
		$last_s_idx = $s_idx;
	    }
	    $last_time = $r->[TIME];
	}
	# Replace records by sample stats
	$chains{$chain} = \@samples;
    }

    # Return an hash of array of hash.
    return $self->{sample_report} = \%chains;
}

1;

__END__