Binary zeros would not be good for disguising an underlying pattern. Back in the late '60s, the intelligence community could read several layers deep, no matter what had been written over it, so the real ability to recover data today must be almost frightening. I think that with CYCLES(n), a different pattern is written each cycle. If not, it should be.
Regards, Richard Schuh -----Original Message----- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Alan Altmark Sent: Friday, June 23, 2006 3:17 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: DASD Secure Erase On Friday, 06/23/2006 at 04:16 AST, "Lewis, David (SCI TW)" <[EMAIL PROTECTED]> wrote: > With today's DASD the only way to ensure that the data is gone is to write > garbage to cylinder after cylinder, volume after volume, until the device can > no longer store anymore data. Look at ICKDSF TRKFMT FROMRANGE(0,0) ERASEDATA CYCLES(n). It writes a "special pattern" (as opposed to binary zeroes, I guess). Maybe it's "good enough". A company's data policy should indicate how good the erasure must be. Alan Altmark z/VM Development IBM Endicott