Vijaykanth Sadhasivam 990021674
Critique onPaper 6: Ouzzani, M.; Bouguettaya, A Efficient access to Web services, 8(2), 2004, IEEE Internet
Summary:
In this paper, the author describes a query optimization model based on aggregating the quality of web services (QoWS) parameters of different web services to facilitate efficient access and sharing of web services.
A Query Model for Web Services:
A three level query model consisting of
Query level – a set of interfaces that gives users an interface for formulating and submitting declarative queries directed at web services
Virtual level – web service-like operations a particular application domain typically offers
Concrete level – represents the space of web services offered by the web.
Queries are represented as relations at query level to let users formulate and submit queries in a natural way. Mapping rules defines these queries through virtual operations, which represent functionalities a particular domain typically offers. For each virtual operation the relevant Web service (concrete) operations should be located. Each virtual or concrete operation is semantically represented through its Function and Category. There are different levels of matching between the virtual and concrete operations.
Query Optimization:
There should be appropriate criteria for selecting the best service execution plan using QoWS parameters such as latency, fees and availability. The proposed system would monitor these parameters and provides a rating system based on QoWS distance, which is the measured gap between the promised and delivered QoWS.The authors then describe an objective function F which needs to be maximized for each concrete operation involved in resolving a query. Data level optimization consists of ordering of operations invocations, data routing, and data manipulation operations for collecting results and is required to ensure that the obtained ordering is feasible.
The authors then propose an algorithm that builds efficient service execution plans and consists of three phases
Initialization and query unfolding – collects all bound variables the algorithm obtains from the query and initializes the service execution plan to an empty set.
Virtual operations ordering – iteratively select the virtual operations that it can invoke at a certain time in the service execution plan.
Service discovery and operations matching – to replace virtual operations with concrete operations to obtain a feasible sequence .The algorithm selects the concrete operation with the highest value for the objective function described earlier.
The author then describes an implementation of the query approach as a service query engine which consists of the following components – Service Locator, Operation matchmaker, Monitoring agent, Query optimizer and execution engine
CRITICAL QUESTIONS:
1)How do we get the various business organizations in a domain to agree on a common ontology for describing concrete web services and virtual operations? What are the characteristics and standards of such anontology?
2)How would the system deal with the dynamic and volatile nature of web services?
What adaptive techniques can be incorporated and what would be their effect on the
replies generated by the system?
3)How effective is the Objective Function in rating the web-services, so that all results of
the queries are valid and useful to the user