Intel has begun developing rootkit-detection technology with a view to preventing malicious code from modifying the host system's memory.

According to an IT Observer report, the system will monitor the OS and other software that could be covertly modified by a rootkit to hide its presence and any security holes it has created. The monitor is implemented in hardware and checks for unnecessary changes made to memory containing system and application code.

Non-running rootkit code will continue to be the responsibility of anti-virus software, Intel said, with the hardware system present to prevent the activated code from damaging the host.

According to the report, Intel's researchers hope to implement the technology in commercially available silicon in the 2008-2009 timeframe. By then Intel should have shipped its 'LaGrande' security system, which is geared to encrypting information held in keyboard and graphics buffers, and to isolating blocks of memory to prevent processes snooping on each other.

As it stands, LaGrande is about protected data and code from outside threats - it's less able to deal with malicious code operating from within the sealed environment. It may be able to tell the user his or her system has been compromised, but not necessarily prevent the attack in the first place.

This is where the new technology, which will presumably be part of LaGrande 2, comes in.

"We need to connect the computers directly to the data, so the human beings don't have to be the I/O channel, and elevate the role of the human being to a more supervisory role," said Intel's director of research, David Tennenhouse.

LaGrande uses a Trusted Platform Modile (TPM) chip connected to the chipset. ®