If you lines are sorted by it's ID in the file, then use binary search, overvise try:

#!/usr/bin/perl -nw
use constant ID => 'id_tag'; # ID what you searching
use constant ID_LENGTH => length(ID);

if (substr($_, 0, ID_LENGTH) eq ID ) {
  print "found:\n$_";
  exit;
}
[download]

Update: usage: script.pl BIG_FILE.txt

please tell how fast my code on your data, thanks

Update: If your big file contains constant data, then you may create an index file. The format of that file is :

# [id_tag] [byte offset]
id_tag1 0
id_tag2 123
id_tag3 463
...
[download]

Thus you are able to search a tag in the index file and fetch searched line using seek