EET 493

Design for Reliability and

Testability of Digital Systems

Instructor: Dr. Mihaela Radu

The World of Dependability

“If anything can go wrong, it will”

Murphy’ s law
 Dependability

Dependability
=
•The property of a system such that
reliance can justifiably be placed on the
service it delivers
•It is the quality of a service that a
particular system provides.
•Is the ability of system to deliver its
intended level of service to the users.

3
 Dependability

 EN:dependability
 DE:Zuverlässigkeit, Verlässigkeit
 FR: sûreté de fonctionnement
 IT: garanzia di funzionamento
 JP:
 GR: axionistia
 ES: dependencia
 RO: fiabil
 NL: “vertrouwbaarheid”

4
Dependability Tree (author Prasad)
’96) reliability
availability
Attributes/fea safety
tures confidentiality
testability
maintainability
fault prevention
Dependability fault tolerance
Means/ways fault removal
fault forecasting
faults
impairments errors
failures
5
 System Design
(Requirements)

Determine which Dependability attribute

primarily needs to be improved for a
specific application.
“Different goals for different applications”

6
Early computer systems-were based on unreliable
components
 ENIAC, with 17.5K vacuum tubes and 1000s of other
electrical elements, failed once every 2 days (avg. down time
= minutes)

 The first theoretical work in dependable (FT) computing is

credited to John Von Neumann. In 1952, he presented a
series of lectures in the use of replicated logic modules to
improve system’s reliability. Later, he developed an article
entitled:
“ Probabilistic Logics and the Synthesis of Reliable
Organisms from Unreliable Components”
 In this article he presented the concept of majority voting
and analyzed the impact that such arrangements could have
on the probability of a system producing erroneous results.7
 APPLICATIONS

Critical-computation applications
(also known as Safety Critical)
 Critical to human safety
 Aircraft flight control systems,

 Space shuttle

 Requirements/example

Aircraft flight control navigation systems:

Reliability (some hours) = 0.999 999 9 = 0.97 at the end
of 3 hours period (mission time)
 Critical to the environment
 Industrial control systems for a chemical plant, nuclear
plants, etc.
8
 APPLICATIONS
High Availability Applications

Users want to have a high probability of receiving service

when it is requested.
 Transaction processing
 ATM: < 10 hours/year unavailable

 airline reservation: < 1 min/day unavailable

 Telecommunication systems: < 5 min./year unavailable

9
 APPLICATIONS

Long-Life Applications
 Unmanned space flights and satellites; repair is
impossible or prohibitively expensive.
 Examples: Mariner, Explorer and Voyager missions,
satellites

 Typical requirements of long-life application are to

have 95% probability (0.95) to be operational at
the end of mission (e.g. 10 years).
 May be degraded / reconfigured before the end of the
mission time (if operator interaction is possible).

10
 New emergent applications

 Health industry,
 Automotive industry,
 Industrial control systems and production lines,
 Banking, E-commerce,
 Wired and wireless networked applications
 Distributed, networked systems (reliability and security
are the major concerns)

11
 Dependability Tree
Reliability
Availability
attributes Safety
Confidentiality
Testability
Maintainability
fault prevention
dependability fault tolerance
means fault removal
fault forecasting
faults
impairments errors
failures
12
 Reliability
 R(t) = probability that the system is
conform to specifications during the period of
time [t0,t], if OK at t0.

 Reliability R (t) is the probability that the

system produces correct output.

 It is the conditional probability that a system

performs correctly throughout an interval of
time [to, t], given that the system was
performing correctly at time to.

13
 Reliability
 We need systems with high reliability if:
 no temporary deviation from specifications are
allowed (aircraft, heartpace makers):
 aircraft: R(some hours) = 0.999 999 9 = 0.97

 if no repair is possible:
 satellite, spacecraft: R (some years) = 0.95

14
 Reliability Function R(t)
Exponential Failure Law
 t
R (t )  e
Where: -constant failure rate

1
0.8
0.6
0.4 R

0.2
0
time

15
 Availability

 A(t) = probability that the system is conform to

specification at time t.

 Availability A (t) is a function of time.

A(t) is defined as the probability that a system is
operating correctly and is available to perform its
functions at the instant of time, t.

Depends on: failure rate-λ, repair rate-μ, related to

MTTF (mean time to failure) and MTTR (mean time to
repair).

 % of time that a system is available

 A = MTTF/(MTTF+MTTR)
16
 Availability
 High availability of a system means:
 high reliability, or

 failing often and repaired fast

Steady –State Availability- calculated as down time per year

 Example: Ass=90% (36.5 days/year)

 Examples
 Transaction processing
 ATM: Ass=0.93 (< 10 hours/year unavailable)
 Embedded

 telecom: Ass=0.95 (< 5 min./year unavailable)

17
 Safety
 S(t) = probability that a system is either conform to
specification, or stopped in a safe way, at time t.

 Safety can be improved through the incorporation of

features that provide the fail-safe operation of the
system. For example, if a fault occurs, a mechanism
must be provided to detect the existence of the fault and
to prevent the fault from resulting in an undesired
response from the control system.

 In essence, safety is the ability (or a measure) of a

system to be fail-safe.

 Safety is a probability that safety actions will occur.

18
Safety requirements for aircraft

19
 Reliability versus Safety

 Reliability  Safety
 Provide correct service  Avoidance of hazards
 Incorporates functional  Incorporates non-
specifications functional specifications
 No severe consequences  May have catastrophic
when it fails consequences when it
fails

different costs of failure

different methodologies for system construction
safety =reliability

20
 Maintainability

 M(t) = probability that the system is

back to specifications at time t
if failed at t0.

 How fast can be repaired, once it failed?

 Restoration process:
 Locating the problem. Physically repairing the
system. Bringing to its operational condition.
 Automatic diagnosis

21
 Testability- Related to Maintainability

 Testability
 ability to test the
characteristics of the
system

 Example:
 Design for Testability of
Digital System:
 BIST….

22
 Design for Testability
 Testability –a design characteristics that influences
various costs associated with testing.

 Design for testability techniques are design efforts

specifically employed to ensure that a device is testable.

Attributes: Controllability and Observability

 Controllability is the ability to establish a specific signal
value at each node in a circuit by setting values on the
circuit’s inputs.
Observability is the ability to the determine the signal
value at any node in a circuit by controlling the circuit’s
inputs and observing the outputs

23
Other Attributes of Dependability
 Confidentiality (security)
 Non-occurrence of unauthorised disclosure of
information

 Integrity
 Non-occurrence of improper alterations of
information

24
…. Conclusions:
Dependable Systems

Fault, Error,
Reliability
Availability Failures
Safety

Fault Testing and

Tolerance testability

25