Está en la página 1de 23

VOLARE : ADAPTIVE WEB SERVICE DISCOVERY IN MOBILE SYSTEMS

Panagiotis Papakos, David S Rosenblum, Licia Capra, Arun Mukhija


University College London

CAMPUS 2009 12/06/09

Project Overview motivation

Introduction The VOLARE Middleware Solution Example Scenario Objectives and Novelties Implementation and Evaluation Conclusions

Project Introduction motivation


Background
Web Services are becoming increasingly popular. Smartphones make more complex functionality feasible. Extensive work on applications using web services in the mobile world. The use of web services in the mobile environment presents unique challenges :

Resources are limited. The context of a mobile device, and therefore its requirements and
capabilities, constantly change during runtime. An adaptive approach is required.

Project Introduction motivation


Typical Web Service Discovery
A provided web service will generally be accompanied by a semantic description, detailing the QoS levels offered etc. A client device can make a web service request that includes QoS levels that need to be satisfied by potential providers. These semantic descriptions and requests use Ontologies to ensure common term usage. A third party Broker will match the service requests with the appropriate web services from a repository

Project Introduction motivation


Service Requesting
Typical Web Service broker based service discovery with a mobile device as a client
Service Request Query

Client Device
Service Binding

Broker
Matching Services

Service Repository

Problem
The web service requirements (QoS requirements, functionality etc) of a mobile device may change due to changes in the context or the resources of the mobile device. Context : Location, Speed, Other Local Devices, User Preferences etc. Resources : Battery Level, Memory Usage, CPU load etc
5

Project Introduction motivation


Service Provision
Typical Web Service broker based service discovery with a mobile device as a service provider
Provider Device
Service Description

Service Repository

Query

Broker
Matching Services

Problem
The web service provider capabilities of a mobile device may change depending on its resources and context. This can lead to bottlenecking. A method to adapt the web service requests and descriptions to reflect the current needs and capabilities of a mobile device is needed.

Project motivation
Proposal
Mobile Device

Context and Resource data Service Request

VOLARE Middleware

Adapted Service Request

Query

Broker
Matching Services

Service Binding

A middleware that : Monitors the Resources of the Device. Monitors the Context of the Device. Dynamically adapts the web service Request appropriately. Dynamically adapts the web service description of the provided services accordingly. Rebinds web services when old bindings no longer satisfy requirements.
7

Project motivation Overview

Introduction The VOLARE Middleware Solution Example Scenario Objectives and Novelties Implementation and Evaluation Conclusions

Project The VOLARE motivation Middleware Solution


Prototype details
An adaptation policy language under

development for VOLARE.


Adaptation policy files used to describe the

relationships between context or resource variables and QoS metrics in service requests/descriptions. Service Descriptions and adaptation policies separate. This allows implementation of VOLARE without changing existing service descriptions. Compatible with non-VOLARE service providers/clients.

Allows adaptation of variables not included in the service descriptions (i.e trust). 9

Project The VOLARE motivation Middleware Solution


Further Issues
Rebinding of Web Services A method needs to be developed for evaluating when it is desirable to rebind to another service.
On-the-fly deployment and unemployment of web services. Balancing between Client and Provider functionalities, in the case of mobile devices that act as both. By altering the QoS levels offered or requested, the two functionalities can be balanced according to the current needs of the device.

10

Project motivation Overview

Introduction The VOLARE Middleware Solution Example Scenario Objectives and Novelties Implementation and Evaluation Conclusions

11

Project motivation Example Scenario


The Scenario Traffic Mapping Software
One possible application of VOLARE is in Traffic Mapping: Assume a fleet of cars equipped with GPS guidance devices. These devices use web services to : Request traffic data for the surrounding area. Send local traffic data to surrounding devices. The QoS metrics of the traffic information web service is: The refresh rate, with which the traffic information is updated. The bandwidth cost, which will be higher if the refresh rate is higher.

12

Project motivation Example Scenario


1. Normal Operation
As the vehicle drives around, the mobile device will :

Request traffic information so that it can plan the optimal route


Provide local traffic information to surrounding devices at a certain refresh rate

2. Low Traffic
When the vehicle enters a Low Traffic zone, its speed will increase. The device will : Increase the refresh rate it requires. The device must quickly acquire up-to-date surrounding traffic information, to quickly plot the optimal route. since its location is rapidly changing. Decrease the traffic information provided, to conserve resources for receiving traffic updates and rapid route plotting.
13

Project motivation Example Scenario


3. Heavy Traffic
When the vehicle now enters a heavy traffic zone, its speed will decrease. The device will : Decrease the refresh rate it requires, to conserve resources. Since the location changes slowly, there is not need for rapid updates. Increase the refresh rate it provides, since there are more available resources, and it becomes more critical to update incoming cars of its heavy traffic state, so they can avoid the area.

14

Project motivation Example Scenario


Adaptation Policy
For this example, the service description/request would also be attached with an adaptation policy, specifying the relationships between context and resource variables and the QoS metrics of the service description/request.
<service-description> <name>TrafficService</name> <wsdl>www.trafficServices.com/Traffic Service</wsdl> <subject>Traffic Information</subject> <bandwidth> 50 </bandwidth> <refreshrate> 5 </refreshrate> <comments> This Web service transmits the current traffic status of the mobile device when driving.</comments> </service-description>
<adaptation-policy> <resources> Battery, CPU, NetworkAvaliability </resources> <context> Date, Time, Location, Status </context> <adaptation> bandwidth = bandwidth*NetworkAvaliability <status value="Heavy Traffic"> refreshrate = refreshrate*(CPU+Battery)*1.5</status> <status value="Low Traffic"> refreshrate = refreshrate*(CPU+Battery)*0.5 </status> </adaptation> </adaptation-policy>

15

Project motivation Overview

Introduction The VOLARE Middleware Solution Example Scenario Objectives and Novelties Implementation and Future Issues Conclusions

16

Project motivation Objectives and Novelties


VOLARE Objectives
More accurate service discovery in mobile systems.
Improved resource efficiency for Service Centric Systems(SCS), by ensuring: That the service bound can satisfy the current needs of the SCS That the services bound are not costing more than necessary in resources given the requirements of the system. Reliable and resource efficient service provider functionality for mobile devices. Allow mobile devices to act as both clients and providers without compromising their functionality.
17

Project motivation Objectives and Novelties


VOLARE Novelties
No similar adaptation approaches currently applied on the realm of mobile web service discovery.
This project will allow us to evaluate the effects of fully dynamic adaptation in mobile web service discovery. Very little existing work on adaptive web service provider functionality on mobile devices. VOLARE can be used on existing web services without modifying them.

18

Project motivation

Introduction The VOLARE Middleware Solution Example Scenario Objectives and Novelties Implementation and Evaluation Conclusions

19

Project motivation Implementation and Evaluation


Implementation
A series of two prototypes will be implemented to demonstrate the benefits of adaptation in service discovery in mobile systems. 1. The first prototype under development is a proof-of-concept implementation, deployed on actual mobile device to prove the viability of the architecture. 2. The second prototype will be developed as a simulation to allow more controlled evaluation of the effects of adaptation in service discovery.

Evaluation
Evaluation after every prototype iteration, using : Self developed case scenarios. Simulation using context and resource usage datasets. Variables evaluated : 1. Discovery Speed 2. Discovery Accuracy 3. Resource Efficiency
20

Project motivation

Introduction The VOLARE Middleware Solution Case Example Objectives and Novelties Implementation and Evaluation Conclusions

21

Project motivation Conclusions

Use of web services by mobile devices can be problematic, because requirements change according to changes in context and resources. Service provider capabilities also change according to context and resources The VOLARE middleware will adapt the service requests and descriptions according the current condition on the mobile device. Introducing dynamic adaptation in mobile web service discovery will enable: More accurate service bindings. More resource efficient SOAs. More reliable service provision using mobile devices.
22

Project motivation VOLARE

Questions ?

23

También podría gustarte