RASS management Reliability Availability Serviceability and Security By

RASS management Reliability, Availability, Serviceability and Security By Anand Tikotekar

Abstract • Argues and furthers the importance of RASS in software systems (emphasis on Distributed systems) • Central to software development • Existing work done in areas of RASS/RAS • Necessity of all 4 parameters.

Introduction • Reliability: consistency of the results often meeting or exceeding specifications. • Availability: The probability that a system is up and kicking at any given point t. • Serviceability: The ease with which Preventive and corrective maintenance can be performed on the system • Security: State of being free from danger or injury

Importance of RASS management • Distributed systems as opposed to standalone • Loss of Reliability • Lack of Availability • Disruption in Serviceability • Breach in security Consequences of the above.

RASS Vs functionality • Systems can survive with RASS and without extra functionality • Opposite claim is not true • Survivability encompasses RASS • Dependability and Survivability interchangeable

RASS as a process in Software development • RASS depends and guides much of the choices that are made during software life cycle. • RASS as a “Basic Need” in the hierarchy of needs • RASS parameters are decided and evaluated from inception to deployment. • Significant damage if RASS altered after the product is built.

Existing RASS/RAS work • HA OSCAR • Head node redundancy • Infini. Band Architecture • New industry standard designed for server I/O • IBM’s RS/6000 server S family (Reliability is most significant here) • CISCO MGX 8230: Carrier class reliability (system can be configured for 100% redundancy)

Existing RASS/RAS work contd. . • PCI Express (Standard I/O interconnect) • Reliable protocol architecture • Device level protocol error detection • HP integrity superdome (RAS features are classified into • Failure avoidance (Keep it running) • Quick fault recovery (Fix it Fast) • Wind River’s CIRRUS development platform (includes RASS) • • Leverages proven technology. Enables reprogrammable smart devices. • DSI (Distributed security infrastructure) (includes security) • Use of coherent security framework • Kernel level security enforcement

Four-zero system • • • Either four RASS parameters or zero Any other combination not sustainable. A chain is as strong as its weakest link

Measurement • • Reliability Availability (CTMC) Serviceability Security (degree at best approximated)

A different line of analysis • • • Software reliability standards Context of Distributed systems Common dependability standards Competitive advantages A common ground for comparison

Conclusion • • RASS topmost priority among software corps RASS as a force for market share RASS to guide software development RASS as “Readily Added System Services” • R A S S

References • Chokchai Leangsuksun “Availability Prediction and Modeling of High Availability OSCAR Cluster”, IEEE Cluster 2003, Hong Kong, December 2003. • Chokchai Leangsuksun “Reliability Modeling using UML” , The 2003 International Conference on Software Engineering Research and Practice, Las Vegas, June 2003. • Peter Loscocoo, Stephen Smalley, Patrick Mukelbauer, Ruth Taylor, Jeff Turner, John Farrel “The inevitability of failure: The Flawed Assumption of security in modern computing environments” National security agency (NSA) • Lopez-Benitez “Dependability Modeling and Analysis of Distributed programs” IEEE transactions on Software Engineering, May 1994.