Device Driver Reliability - Advanced Operating Systems and Kernel Applications - page 44

and reduces the likelihood that a faulty driver will
corrupt other processes.

Fraser, K., Hand, S., Neugebauer, R., Pratt, I.,
Warfield, A., & Williamson, M. (2004). Safe
hardware access with the Xen virtual machine
monitor. In Workshop on Operating System and
Architectural Support for the On-Demand IT
Infrastructure .

chapter Summary

Reliability has become a critical challenge for
commodity operating systems. The competitive
pressure on these systems and their huge installed
base, though, prevents the adoption of traditional
fault-tolerance techniques.
This chapter presents a new approach to im-
proving the reliability of operating systems that is
at once efficient and backwards compatible. Rather
than tolerate all possible failures, Nooks targets
the most common failures and thereby improve
reliability at very low cost. In today's commodity
operating systems, device driver failures are the
dominant cause of system failure.
Nooks prevents drivers from forcing either
the OS or applications to restart. It uses hardware
and software techniques to isolate device drivers,
trapping many common faults and permitting ex-
tension recovery. Shadow drivers ensure that the
OS and applications continue to function during
and after recovery. Dynamic driver update ensures
that applications and the OS continue to run when
applying driver updates.
The Nooks system focuses on backward
compatibility . That is, Nooks sacrifices complete
isolation and fault tolerance for compatibility and
transparency with existing kernels and drivers.
Nevertheless, Nooks demonstrates that it is pos-
sible to realize an extremely high level of operating
system reliability with low performance lost for
common device drivers.

Herder, J. N., Bos, H., Gras, B., Homburg, P.,
& Tanenbaum, A. S. (2006). Minix 3: a highly
reliable, self-repairing operating system. ACM
Operating Systems Review , 40 (3), 80-89.
doi:10.1145/1151374.1151391

Herder, J. N., Bos, H., Gras, B., Homburg, P., &
Tanenbaum, A. S. (2007). Failure resilience for
device drivers. In The 37th Annual IEEE/IFIP
International Conference on Dependable Systems
and Networks , (pp. 41-50).

Hunt, G., & Larus, J. (2007). Singularity: Rethink-
ing the software stack. Operating Systems Review ,
41 (2), 37-49. doi:10.1145/1243418.1243424

Koch, H.-J. (2008). The Userspace I/O HOWTO.
Revision 0.5. In Linux kernel DocBook docu-
mentation .

Microsoft (2006). Architecture of the user-mode
driver framework . Version 0.7. Redmond, WA:
Author.

Spear, M., Roeder, T., Hodson, O., Hunt, G., &
Levi, S. (2006). Solving the starting problem:
Device drivers as self-describing artifacts. In
Proceedings of the 2006 EuroSys Conference ,
pages 45-58.

Swift, M., Annamalau, M., Bershad, B. N., &
Levy, H. M. (2006). Recovering device drivers.
ACM Transactions on Computer Systems , 24 (4).
doi:10.1145/1189256.1189257

referenceS

Swift, M. M., Bershad, B. N., & Levy, H. M.
(2005). Improving the reliability of commodity op-
erating systems. ACM Transactions on Computer
Systems , 23 (1). doi:10.1145/1047915.1047919

Barham, P., Dragovic, B., Fraser, K., Hand, S.,
Harris, T., Ho, A., et al. (2003). Xen and the art
of virtualization. In Proceedings of the 19th ACM
Symposium on Operating Systems Principles .