Yes, even you can use CPAN. You can install perl modules in your home directory, or into your build tree (along with your code). Either that, or you can't even install your code so the question is moot.

Now that you CAN use modules, you may want Text::CSV_XS to handle tab-separated data (though it may be overkill at this point, I always like planning for as much flexibility as possible, and since I don't need to write the code but merely call it, all the better). But you almost definitely DO want Date::Parse for parsing those strings into unix-times that you can use for simple comparison.

This is a solution regexp based. You should complete the %months hash to include all the other months.

use strict;
use warnings;

my @dates = <DATA>;

my %months = ('Jan' => 1,
              'Feb' => 2,
              'Dec' => 12,
             );

foreach my $str_date (@dates) {
    $str_date =~ /^\w*\s*\w*\s*(\w*)\s*(\d*).*(\d{4})$/;
    my($month, $day, $year) = ($1, $2, $3);
    my $date = sprintf("%4d%02d%02d", $year, $months{$month}, $day);

    if($date gt "20080228") {
        print $str_date;
    }
}

__DATA__
usage1    Thu Feb 28 17:30:47 2008
usage2    Fri Feb 29 03:55:22 2008
usage3    Fri Feb 29 04:00:46 2008
usage4    Fri Feb 29 04:10:48 2008
[download]

That outputs

usage2    Fri Feb 29 03:55:22 2008
usage3    Fri Feb 29 04:00:46 2008
usage4    Fri Feb 29 04:10:48 2008
[download]

If you've got it, you could use the groovy "named capture" feature of Perl 5.10 to change that regex to:

    $str_date =~ /^\w*\s*\w*\s*(?<month>\w*)\s*(?<day>\d*).*(?<year>\d
+{4})$/;
[download]

and then the usage to:

    my $date = sprintf("%4d%02d%02d", $+{year}, $months{$+{month}}, $+
+{day});
[download]

Thank you for the suggestion.

i dont have any permission on server to install a module of perl

If it's only the latter, you can always install modules into a local directory. See the documentation about the PREFIX option to perl Makefile.PL.