Today's system's requirements include distributed and net-centric concepts and capabilities. The challenge is to rapidly and accurately distribute large amounts of information to large numbers of nodes, over a variety of transports, account for application level Quality of Service (QoS) all while realizing a decoupling data-centric software architecture. Traditional point-to-point integration technologies and techniques simply don't work in these large distributed environments.

This tutorial introduces the OMG Data Distribution Service (DDS) and highlights the standard's unique capabilities that address and solve these real-time distributed system integration challenges. Specifically, the talk will include an introduction to publish/subscribe concepts, an overview of the specification, DDS data-modeling capabilities, and many of the application level QoS polices and their uses in real-world applications. The tutorial will conclude with a simple demonstration highlighting key concepts covered.
  

The OMG Systems Modeling Language (SysML) is a general-purpose graphical modeling language for specifying, analyzing, designing, and verifying complex systems that may include hardware, software, information, personnel, procedures, and facilities. In particular, the language provides graphical representations with a semantic foundation for modeling system requirements, behavior, structure, and integration with a broad range of engineering analysis. This tutorial provides an introduction to how SysML's extended subset of UML diagram types can address the needs of the systems engineer.
  

This tutorial will explain the key features and mechanisms in the Lightweight CORBA Component Model (CCM) specification, including in-depth examples of applying the OMG Deployment and Configuration Specification for distributed real-time and embedded (DRE) applications. Several examples will be used to demonstrate how these features and mechanisms can simplify DRE application development and integration, showing how to develop CORBA components, how to assemble these components into applications, and how to deploy these applications in the Lightweight CCM run-time environment. Further examples will show how real-time extensions to Lightweight CCM can enable the development of robust, adaptive, and complex DRE applications. By attending this tutorial, DRE system developers will get a comprehensive introduction to the Lightweight CORBA Component Model. Attendees will also learn how to use CCM features and extensions to develop CORBA applications that can meet stringent real-time QoS requirements. This tutorial will also refer to ongoing OMG standards activities to add QoS support to CCM and UML, as well as enhance CCM by integrating it with Model Driven Architecture (MDA) and Model Integration Computing (MIC) tools and techniques.
  

The OMG Data Distribution Service (DDS) for Real-Time systems is often seen as a QoS-enabled real-time messaging technology that provides very low and deterministic communication latency, and high throughput. This perspective, although accurate, is not complete as it misses some of the most powerful, productive, and innovative DDS features. This advanced tutorial will provide an in-depth explanation of the most powerful mechanism provided by DDS, and specifically its support for (1) relational-information modeling, (2) Object/Relational Mapping (namely DDS DLRL), and (3) event filtering and processing. These features will be explained and showcased by relying on real-world examples that will help understand the concepts as well as their applicability. The workshop will complement the advanced DDS features description and explanation with a catalog of patterns describing their applications to solve some of the most challenging design issues in real-time business- and mission-critical distributed systems.
 

Experience using "vanilla" UML in the development of time-critical and resource-critical systems shows that, while useful, it nevertheless lacks some key modeling elements needed in this domain, including quantifiable notions of time and resources, and specific real-time constructs such as task and semaphore. In addition, some aspects of the language's semantics are also not rigorously-enough defined for RT/E systems development. Fortunately (and contrary to an often-expressed opinion) UML's extensibility features can be used to address these issues, allowing the definition of a language profile that standardizes the UML representation of RT/E concepts. The first such standard was the UML Profile for Schedulability, Performance and Time (SPT), adopted by OMG in 2002, which mainly focussed on model-based RT analysis (and especially rate-monotonic analysis for schedulability analysis and layered queuing analysis for performance analysis). OMG has now introduced a new second-generation UML RT/E profile to supercede SPT - the UML profile for Modeling and Analysis of Real-time and Embedded Systems, or MARTE for short. MARTE has a broader scope than its predecessor, and aims to tackle all the activities of both branches of the classical V cycle- modeling, and validation& verification. Its modeling capabilities support both hardware and software aspects of RT/E systems, in order to improve communication between developers, and foster the construction of models that can be used in quantitative analysis of a design's hardware and software characteristics. This tutorial covers both the foundations of model-driven development of RT/E systems, and the use of MARTE for model-driven development.
 

Complex Event Processing (CEP) is an emerging discipline whose technologies allow continuous event processing (i.e. aggregation over time and across events), an area that complements conventional BPM and SOA while providing useful insights into operational real-time data. CEP is currently enjoying rapid adoption in enterprise-scale time-critical systems, such as algorithmic trading in financial markets, fraud detection, airline operations monitoring, and RFID-enabled supply chains. This tutorial introduces CEP concepts and technologies, and relates them to past and current real-time technologies.
 

Dependability and resilience continue to be issues in real-time and embedded distributed systems middleware. It is difficult to accommodate the wide range of desired application-specific policies and mechanisms in a standard set of features of products that reside "between" the application and the underlying Operating System and communications infrastructure. The necessary performance and complexity costs of these features are also of greater concern in often resource-constrained embedded systems or in systems with real-time responsive constraints. The presentations in this session will explore the frontier in two specific applications: adding Lightweight Fault Tolerance to CORBA and defining a security model for DDS.
 

OMG's emerging Lightweight Fault-Tolerance for Distributed Real-Time Systems (LwFT4DRTS) specification is aiming to define a flexible framework for mission critical applications requiring simultaneous fault-tolerant (FT) and real-time (RT) properties by integrating CORBA with external FT mechanisms while maintaining the portability of applications based on existing CORBA middleware. This presentation will evaluate several ways to design and implement interfaces for key modules, including Replica Manager, Forwarding Agent, and Client Proxy, necessary for implementing the LwFT4DRTS specification using the GIOP LOCATION_FORWARD mechanism for redirecting invocations and the interactions between these modules and external FT mechanisms.
 

DDS and CORBA are increasing used together in complex real-time distributed systems with high availability requirements. The many-to-many communications paradigm and flexible QoS policies of DDS support replication of both publishers and subscribers for high availability, while fault tolerance for real-time CORBA is the subject of a current OMG RFP. This presentation describes several roles that DDS can play in providing real-time fault tolerance for the CORBA interfaces in a distributed system. These include replication of application-level object state, server and object group management, and various approaches to routing messages between clients and replicated servers.
 

A consensus has begun to emerge toward a Security Model for DDS which includes: Mandatory Access Control for Secure Domains, Role-Based Access Control for Secure Topics, and centralized policy management. This presentation will discuss development of a secure DDS solution that can be practically deployed, the techniques available for deploying DDS on typical WAN environments, including networks that contain NATs, native support of secured operating systems such as SE Linux, the definition of a pragmatic Role-Based Access Control model to protect DDS Topics and how to enforce the policies using the tools available in the standard PKI, and the impact on performance and real-time predictability that these techniques introduce.
 

Papers in this session address different approaches to managing the deployment of large, complex embedded applications.
 

This presentation focuses on the technological aspects of the integration of a contract based scheduling framework (FRSH) with a component-based framework based on the Lightweight CCM-specification (MyCCM). First presented is the FRSH technology, which is a contract-based resource reservation framework providing a flexible real-time scheduling service. Then the way of integrating FRSH in the MyCCM framework will be explained. This comprises the main approach, the way to represent timing requirements, how to associate components with contracts in the deployment and configuration plans. And finally, the exposure of the incoming tasks and expected results with such a framework.
 

Since its adoption as a standard by the OMG, the Data Distribution Service (DDS) has gained acceptance as an effective middleware for distributed real-time embedded (DRE) systems due to its performance, flexibility, and quality of service (QoS) policies. The DDS standard, however, does not specify interfaces or meta-data to describe and configure these QoS policies at run-time. As a result, this information is either (1) largely application-driven and coded imperatively, which limits flexibility and reusability, or (2) described and configured in a proprietary manner, which limits portability. A promising approach for addressing the issues raised above involves integrating DDS and the CORBA Component Model (CCM), i.e., implementing participants in DDS communication as components (or component assemblies) that communicate via DDS. This discussion presents alternative approaches to integrating DDS with CCM.
 

To build distributed applications using DDS for large scenarios typically involves a big number of individual processes running on many computers spread on the network. In this presentation we propose a new approach for defining, configuring, and deploying DDS applications. Instead of a programmatic approach, our system is based on XML. We show how to define a flexible XML-based language to describe the DDS entities and their associated QoS as well as how they should be deployed on the network. In addition, we will report preliminary tests and lessons learned in using this new methodology.
 

This presentation describes an extension to the OMG's D&C specification (formal/06-04-02) proposed for the design of real-time and distributed component-based applications. It covers all aspects in the deployment process necessary for component- based schedulability analysis. It extends the data models with meta-data for the description, configuration evaluation, and predictions on the applications' temporal behavior. It proposes an architecture for the development process managing repository, as well as the interfaces and the functionality expected for the supporting tools. The extensions have been formulated at the PIM level, independently of the components technology, the platform, and the modeling and real-time analysis methodology.
 

Papers in this session examine new developments in real-time communication protocols.
 

"Data conferencing" applications coordinate continuous data streams. These applications require timeliness, reliability, and multiple senders/receivers. The challenge is to develop COTS-compatible middleware  technologies to support these requirements. We have addressed this challenge by integrating the Data Distribution Service (DDS) with the Ricochet++ transport protocol framework. Ricochet++ provides composable modules for fine-grained transport protocol control as well as dynamic reconfiguration to adapt to changing environments and ensure QoS as needed by the application.

This presentation compares the performance of DDS using various transport protocols including the Ricochet++ transport protocol framework and classifies the protocol properties applicable to data conferencing.
   

DDS is becoming the de-facto standard for distributing real-time publish-subscribe information in performance and QoS-sensitive systems such as Combat Management Systems, Air Traffic Management and SCADA systems. Meanwhile many non-real time systems are built using Web-Service Technologies. These technologies must be integrated. We analyze two approaches to bridging DDS and Web-Service technologies. One based on the definition of a WS-DDS service using WSDL and another mapping DDS directly to HTTP following RESTful principles. We conclude with recommendations and guidance for a potential future OMG standard.
 

The roundtable sessions provide an opportunity for attendees to discuss the workshop's topics in more depth. Attendees will divide into round-table groups, each led by speakers and session chairs, to examine the ramifications of recent technology developments on engineering real-time and embedded systems. Each roundtable will last 45 minutes, followed by a 15 minute plenary summarizing the conclusions to the other workshop attendees.
 

Poster presentations will be available for viewing throughout the day on Tuesday. During this session each poster presenter will give a very short summary of their work. Poster presenters will then be available to discuss their posters during the reception.
 

Formal verification of component-based DRE systemic properties requires developers to understand multiple aspects of the system's behavior and verification techniques that are usually outside their knowledge domain. This poster presents our preliminary results for formally specifying component-based systems for verification of systemic properties, such as worst-case response time. We discuss the challenges of formally specifying component-based systems and show how using domain-specific modeling languages, which operate at a level of abstraction that is more intuitive to DRE system developers, helps simplify formally specifying and, in turn, verifying systemic properties.
   

We present a tool-chain which converts system models into implementation code using model transformation techniques. The tool-chain uses two modeling languages: one for modeling systems (as dataflow graphs with statecharts), and another for modeling imperative computer programs. At its heart are two graph-transformation-based model transforms which convert the periodic/sampled dataflow and event-driven statechart components of a system model into implementation program models. A program model is akin to an abstract-syntax-tree, and so is amenable to validation and verification, as well as direct interpretation into any imperative language such as “C” or Java.
 

In this presentation, we describe Model Transformations Templatization and Specialization (MTS), which is a novel model transformation technique that overcomes the challenges of middleware quality of service (QoS) configuration for heterogeneous distributed, real-time systems by supporting the concept of templatized model transformations and their subsequent specializations. Using MTS the variabilities of each target platform model can easily be captured as models (as opposed to transformation rules), which are later provided as arguments to the templatized transformation for instantiating mappings for that target platform. The principles and techniques supported by MTS can be implemented on existing model transformation tool suites. The presentation will discuss how we have applied MTS in the context of publish/subscribe communication semantics, such as CCM real-time publish/subscribe service.
  

Publish/subscribe is a widely used middleware architecture because it promotes loose coupling between interacting components stemming from the communication semantics, which are inherently anonymous, asynchronous, and multicasting in nature. In spite of these features, the requirements of mobile applications, such as vehicular systems, are hard to satisfy using contemporary publish/subscribe middleware that are developed for static and stationary systems. Frequent connections/disconnections, scarcity of bandwidth and power, and frequent changes in location are the major challenges. These may be coupled with changes in the terrain and potential obstructions by objects, such as buildings, as in the vehicular environment. Nevertheless, there exist applications, such as automated route guidance, that are time critical while requiring the real-time and reliable delivery of information to a large number of vehicles. We discuss why existing publish/subscribe middleware solutions for mobile environments, including DDS, fall short in satisfying the requirements of highly dynamic systems like the vehicular systems. We propose a solution to overcome their limitations using a technique called restricted flooding that yields higher throughput at the expense of higher messaging overhead in the networking infrastructure.
 

We present our planned system for audio and video transmission over DDS. We aim to build a modular, scalable and extensible system which takes advantage of the QoS DDS policies to solve the usual problems experimented in traditional systems (HTTP or RTP based) for multi-conference applications. Our "data-centric" approach uses the publish/subscribe paradigm to decentralize the information. As a result, better scalability, robustness and flexibility are got compared to conventional systems based on the client/server paradigm. Our system uses the DDS middleware for directory service provision and for data transmission and signaling. Our work shows that audio, video, and signaling transmission over DDS is not just possible but appropriate.
 

Developing distributed applications is a difficult and complex process due, in part, to excessively feature rich frameworks, the need to minimize resource utilization, and the need to adapt to changing conditions. In order to address these problems, traditional middleware must evolve through specialization techniques. The specialization provided by distributed middleware plays a vital role in determining the end-to-end quality of service (e.g., performance, footprint) offered by the distributed applications based on them. Specialization can be though of as a combination of optimization before runtime and adaption during runtime. This mandates a need for a systematic, high-level process for enabling specializations at every stage of the application development life cycle. We present a three-dimensional taxonomy that categorizes different middleware specialization approaches, which can serve as a guideline for specializing middleware-based applications in a systematic way.
 

Complex Event Processing (CEP) is the emerging discipline of identifying and acting upon complex groups of timed events. CEP is currently enjoying rapid adoption in enterprise-scale time-critical systems. Papers in this session will discuss not only the application of CEP in commerce, but also its application to large-scale real-time and embedded systems.
 

Businesses today are increasingly "real-time". Conventional IT systems, even when designed around BPM and SOA, are traceable to their predecessors in the era of manual offices: message tubes, card index systems, and "rubber stamps". Different approaches, such as Event Driven Architectures and Complex Event Processing, are proving very useful in replacing or augmenting conventional IT, and can handle the needs of real time industrial application areas like RFID and eXtreme Transaction Processing. This session looks at some of the CEP application requirements in commercial applications and how they are addressed by CEP, using techniques refined from the military sensor fusion world.
 

Within current Naval combat systems, large volumes of high-rate data are generated, aggregated, correlated, assessed, and acted upon at run-time. Existing commercially developed Complex Event Processing (CEP) technologies provide capabilities that can potentially automate (in part) and simplify these tasks. The presentation provides an overview and initial results from NSWCDD's ongoing evaluation of CEP technologies.
 

This presentation describes the results and architecture of a DoD-funded research effort to develop a common normalized information picture combining data from System Monitoring Tools and Intrusion Detection Systems. The normalized informational picture is built using two components: the Data-Distribution Service (DDS) and Complex Event Processing (CEP). DDS was selected because it provides a Standard API and Network Protocol able to handle large volumes of real-time information and prioritize it by setting QoS policies. CEP was selected because it provides a familiar (SQL-like), powerful and extensible language able to process large amounts of streaming data, aggregate the information, correlate it, and uncover interesting events and threats. The purpose of this work is to increase the speed and effectiveness of technologies used to counter network attacks.
 

Papers in this session address different aspects of the problem of measuring the performance of complex, large-scale distributed systems to ensure they meet their QoS & performance requirements.
 

This presentation will describe the results of our work in defining a serialization format to realize a flexible communication. The results of systematic benchmarks will be used to evaluate the pros and cons of various flexible data representation formats available for use with DDS in terms of their latency, serialization, and de/serialization performance. Moreover, several compression techniques are also evaluated in order to alleviate the overhead incurred by these flexible formats reducing the size of data sent through a network.
 

With the growing adoption of DDS as a crucial building block for creating business- and mission-critical systems, there is an increasing need to systematically evaluate the performance of DDS implementations in a variety of environments with a range of application workloads. This requires an open, vendor-independent DDS benchmarking suite.

This talk will cover several aspects of such a suite. First, we will identify operational definitions of performance for DDS-based data-centric systems. Then we will outline the purpose and scope of a systematic DDS benchmark suite. Finally, we will present an overall architecture for realizing the suite.
 

In large-scale distributed real-time and embedded (DRE) systems application-level components are often not developed until after infrastructure components are complete. Unfortunately, this means that developers only realize systems will not meet their performance requirements during system integration. Although system execution modeling (SEM) tools help address this problem, they have not historically focused on efficient testing. This presentation shows how combining SEM tools with continuous integration environments helps improve testing capabilities for DRE systems. It also shows how combining continuous integration environments and SEM tools can help evaluate system performance continuously throughout the development lifecycle.
 

Papers in this session cover the application of Model-Driven Development (MDD) techniques, including Model-Driven Architecture (MDA), to building correct, reliable and predictable real-time and embedded systems.
 

In a recent NATO study, the OMG Model Driven Architecture approach has been used for modelling a standardized architecture for military aircraft, in which requirements for safety, reliability and upgradeability need to be explicitly addressed, along with the need to incorporate pre-existing and upcoming standards. This presentation will report on experiences so far, using examples based on this ongoing NATO standardization initiative. It will illustrate techniques used to achieve not only execution platform independence, which is of course the essence of MDA, but also vehicle platform independence, allowing the models to be deployed on a wide range of aircraft.
 

The UML profile for MARTE provides concepts for modeling qualitative and quantitative features of real-time and embedded applications at a high abstraction level. It defines precise semantics for these concepts, and makes them exploitable to build executable models. The presentation first gives an overview of the MARTE concepts relevant for model execution, focusing on the Model of Computation and Communication sub-profile. Then, a modeling methodology for producing MARTE executable models and an associated execution framework are presented. Finally, a mapping of MARTE-based models onto the execution framework is shown.
 

Ericsson has been and is very successful in the 3G (WCDMA) Mobile Network Systems market and has about 50% market share. One of the reasons behind this success is the way these systems have been developed, using Model Driven Development. This presentation describes the experiences from using MDD in development of complex embedded real-time systems and hence in large projects. The presentation starts with a background describing the 3G history, technology, how the systems are currently developed and lessons learnt. Furthermore, the presentation will discuss where Ericsson is heading in the future with MDD including executable and re targetable action language.
 

To set up a communication over DDS you traditionally define (a) the data types, (b) the topic name(s), and (c) the quality of service (QoS) settings to be used. Data types are typically defined in IDL while topic names and QoS are hardcoded. Now, we don't like hardcoded configuration data! Instead, using Eclipse EMF, we have defined a metamodel which brings all these concepts together. At runtime we interpret instances of this metamodel, and dynamically register types with the DDS runtime. Support classes prepare participants, publishers, subscribers, and the lot. We will discuss possibilities and limitations using these ideas.