www.s-cube-network.eu S-Cube Learning Package Service Level Agreements: Analysis Operations on SLAs: Detecting and Explaining Conflicting SLAs ISA Group, University of Seville (Spain) C. Müller, O. Martín-Díaz, A. Ruiz-Cortés, M. Resinas
Transcript
1. S-Cube Learning Package Service Level Agreements: Analysis
Operations on SLAs:Detecting and Explaining Conflicting SLAs ISA
Group, University of Seville (Spain)C. Mller, O. Martn-Daz, A.
Ruiz-Corts, M. Resinas www.s-cube-network.eu
2. Learning Package Categorization S-Cube Quality Definition,
Negotiation and Assurance Quality Management and Prediction
Analysis Operations on SLAs: Detecting and Explaining Conflicting
SLAs ISA Group
3. Learning Package Overview Problem Description Problem
Approach Solution Approach Solution Validation Discussion
Conclusions ISA Group
4. Problem Description:A Logistic Scenario Package Tracking
Provider
5. Problem Description: A Logistic ScenarioTemplate-Offer based
mechanism to create SLAs. An example of two van companies offering
different QoS Package Tracking Provider
6. Problem Description: Temporal-Aware SLAs in the
ScenarioTemporal-Aware SLAs are createdwhen temporal-aware
documents are used.In the example two providers sendagreement
offers with different QoS indifferent time periods.
7. Problem Description: Detecting Conflicting SLAs Within
documents At any SLA lifecycle phase. Is my Offer (for instance, at
edition time) Conflict free? Is my Template Conflict free? Is the
sent Offer compliant with the Template? Is the received Offer
compliant with my Between documents Template? At any SLA lifecycle
phase in which documents compliace[Answer for Questions: must be
assured. YES or NOT] (for instance, at negotiation time)
8. Problem Description: Explaining Conflicting SLAs Within
documents At any SLA lifecycle phase. Why is conflicting (for
instance, at edition time) my Offer? Why is conflicting my
Template? Why Is the sent Offer non-compliant with the Template?
Why is the received Offer non-compliant Between documents with my
Template? At any SLA lifecycle phase in which documents
compliace[Answer for Questions: must be assured. Conflicts
Explanations] (for instance, at negotiation time)
9. Problem Description:WS-Agreement as Case StudyWS-Agreement -
Definition: An XML language and a protocol for Advertising the
capabilities of service providers in templates Creating agreements
based on creational offers and templates I want an agreement *
Service Features: * Guarantees:
10. Problem Description:WS-Agreement as Case StudyWS-Agreement
Template vs. Offers Differences
11. Problem Description:WS-Agreement as Case StudyWS-Agreement
Template vs. Offers Differences Use case An optional Qualifying
Condition (QC) may be included to enable/disable guarantees.
Methaphor
12. Learning Package Overview Problem Description Problem
Approach Solution Approach Solution Validation Discussion
Conclusions ISA Group
13. Problem Approach:Temporal-Aware WS-Agreement
docs.Temporal-Aware Scenario Mon-Sat (Weekly) From Jan 01/2013 to
Dec 31/2013 Sunday (Weekly) From Jan 01/2013 to Dec 31/2013
14. Problem Approach:Temporal-Aware WS-Agreement
docs.Temporal-Aware Scenario Mon-Sat (Weekly) From Jan 01/2013 to
Dec 31/2013 Sunday (Weekly) From Jan 01/2013 to Dec 31/2013 Mon-Fri
(Weekly) From Jan 01/2013 to Dec 31/2013
15. Problem Approach:Temporal-Aware WS-Agreement
docs.Temporal-Aware Scenario Mon-Sat (Weekly) From Jan 01/2013 to
Dec 31/2013 Sunday (Weekly) From Jan 01/2013 to Dec 31/2013 Mon-Fri
(Weekly) From Jan 01/2013 to Dec 31/2013 Mon-Fri (Weekly) From Jan
01/2013 to Dec 31/2013 Sat-Sund (Weekly) From Jan 01/2013 to Dec
31/2013
16. Problem Approach:Kind of Conflicts to Detect & Explain
Within documents Inconsistencies: Contradiction between terms,
parts of terms, or creation constraints, and all of these amongst
themselves; constitute an inconsistency of the WSAgreement
document. This means that it is impossible to find a satisfactory
assignment to the variables that appear in those terms or creation
constraints. The consequence is that the whole document is unusable
because it will never be fulfilled regardless of the way the
service is provided. Term inconsistencies can be originated by any
combination of service description terms, guarantee terms and
creation constraints, including such elements with themselves
17. Problem Approach:Kind of Conflicts to Detect & Explain
Within documents Inconsistencies: (Sample 1/2) The messenger
company impose a cost formula depending on the package weight. But
the guarantee term involving the cost is inconsistent with such
formula.
18. Problem Approach:Kind of Conflicts to Detect & Explain
Within documents Inconsistencies: (Sample 2/2) The messenger
company assure a wrong DeliveryTime by mistake.
19. Problem Approach:Kind of Conflicts to Detect & Explain
Within documents Dead Terms: This conflict is caused when the
condition of a conditional term can never be satisfied in a
WSAgreement document. In other words, a dead term is a guarantee
term whose qualifying condition has a contradiction with itself or
one or more terms and/or creation constraints of the WSAgreement
document. Consequently, this kind of conflict makes the term dead
because its service level objective (SLO) can never be applied
since its precondition never holds.
20. Problem Approach:Kind of Conflicts to Detect & Explain
Within documents Dead Terms: (Sample 1/1) The messenger company
wanted to assure a maximum delivery time for the express and
premium delivery types. But there was a wrong multiple value
assignment in the QC, by a mistake while mapping the natural
language condition to a WS-Agreement condition. The right condition
would be: DeliveryType = Premum OR DeliveryType = Express
21. Problem Approach:Kind of Conflicts to Detect & Explain
Within documents Ludicrous Terms: This conflict is caused when a
conditional term makes the document inconsistent if its condition
holds, which contradicts usual expectations. In other words, a
ludicrous term is a guarantee term with the following
characteristics: 1. It is not inconsistent 2. Its qualifying
condition holds 3. Its SLO does not hold because of a contradiction
within itself, with the qualifying condition or with other terms or
creation constraints of the WSAgreement document. Consequently, the
ludicrous term is one that when the qualifying condition holds, the
SLO and, hence, the guarantee term, does not hold.
22. Problem Approach:Kind of Conflicts to Detect & Explain
Within documents Ludicrous Terms: (Sample 1/1) The messenger
company wanted to assure a maximum cost for deleveries of 1 kg
packages and the delivery types express and premium. In this case
the QC is correctly defined but when its fulfills the SLO cannot be
satisfied due to the cost formula depending on weight.
23. Problem Approach:Kind of Conflicts to Detect & Explain
Within Temporal-Aware documents Inconsistencies, dead and ludicrous
terms conflicts applies but only if the involved terms or creation
constraints are defined in overlapped time periods The sample does
not include any conflict because working and home period
definitions are nor overlapped (see slide 15)
24. Problem Approach:Kind of Conflicts to Detect & Explain
Within Temporal-Aware documents Current sample includes a
inconsistency conflict between locating accuracy guarantees because
working and Christmas period definitions overlaps as can be seen
below
25. Problem Approach:Kind of Conflicts to Detect & Explain
Between documents WS-Agreement Compliance definition The same
agreement responder and template id in both documents Offer service
terms must be compliant with template creation constraints There
must be a compliance between the terms of both documents
26. Problem Approach:Kind of Conflicts to Detect & Explain
Between documents WS-Agreement Compliance definition In current
sample, assuming the same agreement responder and template id, The
offer service terms are compliant with template because Cost has a
right value assignment considering the template cost creation
constraint. But guarantee terms of both documents are not
compliant. Then documents are not compliant.
27. Problem Approach:Kind of Conflicts to Detect & Explain
Between Temporal-Aware documentsPrevious compliancedefinition
applies but only ifthe involved terms or creationconstraints are
defined inoverlapped time periodsThe sample does not includeany
conflict because workingand home period definitionsare not
overlapped(see slide 15)
28. Problem Approach:Kind of Conflicts to Detect & Explain
Between Temporal-Aware documentsCurrent sample includes
anon-compliance betweendelivery time guarantees atworking
period.
29. Problem Approach:Kind of Conflicts to Detect & Explain
Between Temporal-Aware documentsCurrent sample includes
anon-compliance betweenservice term 1 and thetemplate creation
constraints.This situation me beintentional by a try to pay lessfor
the delivery service fromthe client who defined theagreement
offer.
30. Learning Package Overview Problem Description Problem
Approach Solution Approach Solution Validation Discussion
Conclusions ISA Group
31. Problem Approach:Automating Analysis of SLAs There are
several approaches that can be used to detect and explain conflicts
within documents and between documents automatically. For instance,
one may think of creating ad-hoc algorithms to detect the different
types of conflicts. Or one may think of using an ? Analyser
existing paradigm that provides generic algorithms to detect them
such as: Description Logics Constraints Satisfaction Problems
32. Problem Approach:Automating Analysis of SLAs In this
presentation we focus on the use of constraint satisfaction
problems (CSPs) as the paradigm to tackle this problem
automatically. The reasons are: The most significant part of an SLA
is Constraint a set of constraints that are set on some properties
or descriptions of the service and, therefore, CSPs can be ?
Satisfaction Problems used to describe this problem in a very
natural way. There are a plethora of CSP solvers with efficient
algorithms that can be used to detect conflicts automatically. This
approach has been used successfully in many related work.
33. Background:Constraint Satisfaction Problems CSPs allow
modelling and solving expressive problem models based on
constraints and they have been an object of research in Artificial
Intelligence over the last few decades. A CSP is a threetuple of
the form (V , D, C) where V = is a finite set of variables, D = is
a finite set of domains (one for each variable) and C is a set of
constraints defined on V. b The solution of the CSP is the value
assignments to the set of variables V that satisfy all of the
constraints C a S-Cube
34. Background:Constraint Satisfaction Problems For instance,
given the CSP: ({a, b}, { [0, 1], [0, 1] }, { a = b }) A solution
is whatever valid assignment of all elements in V that satisfies C.
b 1 1,1 a, b: integer [0,1] a=b 0,0 0 1 a Operation solve( V, D, C)
is the operation of CSP solvers that returns the solutions of a CSP
and many heuristics and techniques have been developed to obtain
these solutions in an efficient manner. S-Cube
35. Background:Constraint Satisfaction Problems Operation
explain (V, D, C) involves providing an explanation when such
solution is not possible. This explanation is a minimal set of
constraints c C that makes it impossible to find a valid assignment
of all elements in V that satisfies c, i.e., that makes solve(V ,
D, c) = For instance: In the CSP: ({a, b, c}, {[0..2], [0..2],
[0..2]}, {a + b < 1, c = 2}) The explantion c = {(a+b