Review articleModelling web-service uncertainty: The angel/daemon approach☆
Section snippets
Programming in an uncertain world
From the 1940s until the 1970s computations were largely sequential and deterministic. Theories were developed to reason about such computations: Tony Hoare and Edsger Dijkstra constructed a logical basis for establishing the correctness of programs relative to their specifications. At the same time Juris Hartmanis and Richard Karp developed complexity theory for reasoning about the conceptual execution time of programs.
The development of computer networks resulted in an increase in the
Services and orchestrations
Consider a system which utilises a number of web services. Uncertainty can arise in two different ways;
- 1.
failure of system components may give rise to incomplete results.
- 2.
delay of components to respond may cause system slow down.
According to our joint work with Maria Serna, in order to analyse both situations two mathematical models have been devised; a multiset semantics is used to assess system functionality [11] while a “time”-based approach is used to assess system delays [3].
A service is a
Measuring resilience
Resilience is the “ability to be successful again in the face of adversity”; it is a key feature of the behaviour of large-scale web applications. In our work with Maria the resilience of an orchestration under stress is determined by measuring either the quality of its output or its response delay.
The quality of a publication can be measured in many different ways: for example, an image may be assessed by its resolution (a stressed service may publish only a thumb-nail image). In an extreme
Uncertainty profiles and /-games
Uncertainty profiles and / games provide the key techniques for analysing uncertain systems — see for example [1], [2]. We would like to point out that we start our joint work with Maria Serna in 2008 with [2].
Consider (informally) how environmental factors can impact on a large web application. Web environments often change very rapidly, making it difficult, if not impossible, to construct a meaningful record of historical events and their associated probabilities. Instead application users
Block matrix multiplication: stress and redundancy
Ever if Maria Serna was not directly involved in the content of this section we would like to present it as an off-spring of our fruitful collaboration.
A distributed matrix product algorithm may be implemented using cloud resources. However, the performance of an application running on a virtual machine in the cloud can be affected both by the way in which resources are allocated at run-time as well as by hardware contention issues [19], [20].
Divide and conquer algorithms decompose large scale
Extension to social applications
The approach combined with a stress model (Definition 5) can be applied to other application areas, including short-term macroeconomic models [9], [26] and voting games [10], [27]. Maria Serna directly worked on both extensions.
Conclusions and open problems
On this pages we have developed our joint work with Maria Serna along 10 years, from 2007 to 2017. Now, we summarise the results and raise some open questions.
Programming has evolved from the development of codes designed to run on sequential machines in controlled environments to the orchestration of a set of remote, globally accessible services. The range of services now available includes news pages, email providers, storage facilities and cloud resources (see, for example, Section 2). The
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Prof. Joaquim Gabarro, was born in Barcelona (Spain) in 1953 and studied Physics and Computer Science, obtained his Ph.D. from Universitat Politècnicade Calatunya in 1983. From 1985 is Associate Professor (Profesor Titular) in Universitat Politècnica de Calatunya. Main research topics are Concurrency, Algorithms, Game Theory and Complexity.
References (44)
- et al.
A timed semantics of orc
Theoret. Comput. Sci.
(2008) - et al.
Bounded site failures: An approach to unreliable grid environments
- et al.
Analysing orchestrations with risk profiles and angel-daemon games
- et al.
Analysing web-orchestrations under stress using uncertainty profiles
Comput. J.
(2014) - et al.
Computational aspects of uncertainty profiles and angel-daemon games
Theory Comput. Syst.
(2014) The Age of Uncertainty
(1977)- et al.
Animal Spirits
(2009) Risk, Uncertainty and Profit
(1921)- et al.
Theory of games and economic behavior
(1953) - et al.
Uncertainty in basic short-term macroeconomic models with angel-daemon games
IJDATS
(2017)
An angel-daemon approach to assess the uncertainty in the power of a collectivity to act
The computational complexity of qos measures for orchestrations
J. Comb. Optim.
Computation orchestration: A basis for wide-area computing
Softw. Syst. Model.
The orc programming language
A tree semantics of an orchestration language
Reasoning about orchestrations of web services using partial correctness
Form. Asp. Comput.
Abstraction, Refinement and Proof for Probabilistic Systems
Communicating Sequential Processes
Orchestrating high performance services : theory and practice
Uncertainty in the cloud: An angel-daemon approach to modelling performance
An Introductions to Game Theory
Computational Complexity
Cited by (1)
A review on the applications of bayesian network in web service
2023, Research Square
Prof. Joaquim Gabarro, was born in Barcelona (Spain) in 1953 and studied Physics and Computer Science, obtained his Ph.D. from Universitat Politècnicade Calatunya in 1983. From 1985 is Associate Professor (Profesor Titular) in Universitat Politècnica de Calatunya. Main research topics are Concurrency, Algorithms, Game Theory and Complexity.
Dr H.A. Stewart, was born in Portadown (U.K.) in 1960 and studied Mathematics and Computer Science at The Queen’s University of Belfast. He obtained his Ph.D. (also Q.U.B.) in 1986. From 1986 until 2016 he was a lecturer in Computer Science at Q.U.B. His main research interests are: program correctness; parallelism; semantics; parallel algorithms; web-based computation; game theory.
- ☆
This work was partially supported by MINECO, Spain and FEDER, Spain funds under grant GRAMM (TIN2017-86727-C2-1-R) and Generalitat de Catalunya, Agència de Gestió d’Ajuts Universitaris i de Recerca, Spain , under project 2017 SGR 786 (ALBCOM).