You must have disks fast enough to store 2G and 200,000 articles per day.
Usually the trouble is in the number of articles. The operating system and disks must be able to support adding 140 new files per minute: this must include the directory accesses, file creation, write, and file closes just to keep up with the feed. And obviously they must be deleted as fast (on average.) Add client load on top of this.
The FAQ has this to say:
7.20 INN FAQ I am not getting all the articles, but my feeder is sending a full feed
Disk IO is probably the biggest news bottleneck. Usually a full feed
is more than a single Fast SCSI II disk can handle. Having 2 or more disks
for the spool is suggested in either a striped configuration (using ODS
for Solaris or MD for Linux) or a split spool. It is also recomended to
have the disks spread out over multiple controllers.