Section: New Results

QoS for Emergent Mobile Systems

Participants : Georgios Bouloukakis, Nikolaos Georgantas, Rachit Agarwal, Valérie Issarny, Raphael de Aquino Gomes.

With the emergence of Future Internet applications that connect web services, sensor-actuator networks and service feeds into open, dynamic, mobile choreographies, heterogeneity support of interaction paradigms is of critical importance. Heterogeneous interactions can be abstractly represented by client-server, publish/subscribe, tuple space and data streaming middleware connectors that are interconnected via bridging mechanisms providing interoperability among the choreography peers. We make use of the eVolution Service Bus (VSB) (see §  6.2 ) as the connector enabling interoperability among heterogeneous choreography participants. VSB models interactions among peers through generic post and get operations that represent peer behavior with varying time/space coupling.

Within this context, we study end-to-end Quality of Service (QoS) properties of choreographies, where in particular we focus on the effect of middleware interactions on QoS. We consider both homogeneous and heterogeneous (via VSB) interactions. We report in the following our results in three complementary directions:

• While VSB ensures functional interoperability of heterogeneous choreography interactions, differences in timing requirements and constraints of such interactions can severely affect their latencies and success rates. To model timeliness, we introduce the lease and timeout parameters. The former captures data availability and validity in time, while the latter represents intermittent availability of data recipients due to mobility and disconnection. By precisely studying the related timing thresholds using timed automata models, we verify conditions for successful interactions with VSB connectors. Furthermore, we statistically analyze through simulations, the effect of varying lease and timeout periods to ensure higher probabilities of successful interactions. Simulation experiments are compared with experiments run on the VSB implementation testbed to evaluate the accuracy of results. This work can provide application developers with precise design time information when setting these timing thresholds in order to ensure accurate runtime behavior [23] .

• Choreography peers deployed in mobile environments are typically characterized by intermittent connectivity and asynchronous reception of data. In such environments, it is essential to guarantee acceptable levels of timeliness between the data sources and mobile users. In order to provide QoS guarantees in different application scenarios and contexts, it is necessary to model the system performance by incorporating the intermittent connectivity. Queueing Network Models (QNMs) offer a simple modeling environment, which can be used to represent various application scenarios, and provide accurate analytical solutions for performance metrics, such as system response time. We provide an analytical solution regarding the end-to-end response time between the users and the data sources by modeling the intermittent connectivity of mobile users with product-form QNMs. We utilize the publish/subscribe middleware as the underlying communication infrastructure for the mobile users. To represent the subscriber's connections/disconnections, we model and solve analytically an ON/OFF queueing system by applying a mean value approach. Finally, we validate our model using both simulations with real-world workload traces and comparison with an actual implementation of a Java Messaging Service middleware. The deviations between the performance results foreseen by the analytical model and the ones provided by the simulator and the prototype implementation of a real system are shown to be less than 5% for a variety of scenarios.

• Large-scale mobile environments are characterized by, among others, a large number of mobile users, intermittent connectivity and non-homogeneous arrival rate of data to the users, depending on the region's context. Multiple application scenarios in major cities need to address the above situation for the creation of robust mobile systems. Towards this, it is fundamental to enable system designers to tune a communication infrastructure using various parameters depending on the specific context. We take a first step towards enabling an application platform for large-scale information management relying on mobile social crowd-sourcing [26] . To inform the stakeholders of expected loads and costs, we model a large-scale mobile pub/sub system as a queueing network. We introduce additional timing constraints such as (i) mobile user's intermittent connectivity period; and (ii) data validity lifetime period (e.g. that of sensor data). Using our MobileJINQS simulator (http://xsb.inria.fr/d4d#mobilejinqs ), we parameterize our model with realistic input loads derived from the D4D CDR (Call Detail Record) dataset (http://www.d4d.orange.com/en/home ) and varied lifetime periods in order to analyze the effect on response time. This work provides system designers with coarse grain design time information when setting realistic loads and time constraints [18] .