Nima Moghaddami Khalilzad

Resource sharing among prioritized real-time applications on multiprocessors

In this paper, we propose a new protocol for handling resource sharing among prioritized real-time applications composed on a multiprocessor platform. We propose an optimal priority assignment algorithm which assigns unique priorities to the applications based on information in their interfaces. We have performed experimental evaluations to compare the proposed protocol (called MSOS-Priority) to the current state of the art locking protocols under multiprocessor partitioned scheduling, i.e., MPCP, MSRP, FMLP, MSOS, and OMLP. The evaluations show that MSOS-Priority mostly performs significantly better than alternative approaches.

Multi-level adaptive hierarchical scheduling framework for composing real-time systems

Processor partitioning and hierarchical scheduling have been widely used for composing hard real-time systems on a shared hardware platform while preserving the timing requirements of the systems. Due to the safety critical nature of hard real-time systems, a conservative analysis is often used for deriving a sufficient partition size. Applying the exact same analysis for deriving the partition sizes for soft real-time systems result in unnecessary processors overallocation and consequently waste of the CPU resource. In this paper, to address the problem of composing soft and hard real-time systems on a resource constrained shared hardware, we present a multi-level adaptive hierarchical scheduling framework. In our framework, we adapt the processor partition sizes of soft real-time systems according to their need at each time point by on-line monitoring their processor demand. Furthermore, we implement our adaptive framework in the Linux kernel and show the performance of our framework using a case study.

A component based architecture to improve testability, targeted FPGA-based vision systems

FPGA has been used in many robotics projects for real-time image processing. It provides reliable systems with low execution time and simplified timing analysis. Many of these systems take a lot of time in development and testing phases. In some cases, it is not possible to test the system in real environments very often, due to accessibility, availability or cost problems. This paper is the result of a case study on vision systems for two robotics projects in which the vision team consisted of seven students working for six months fulltime on developing and implementing different image algorithms. While FPGA has been used for real-time image processing, some steps have been taken in order to reduce the development and testing phases. The main focus of the project is to integrate different testing methods with FPGA development. It includes a component based solution that uses a two-way communication with a PC controller for system evaluation and testing. Once the data is acquired from the vision board, the system stores it and simulates the same environment that has been captured earlier by feeding back the obtained data to FPGA. This approach addresses and implements a debugging methodology for FPGA based solutions which accelerate the development phase. In order to transfer massive information of images, RMII which is an interface for Ethernet communication, has been investigated and implemented. The provided solution makes changes easier, saves time and solves the problems mentioned earlier.

Towards Energy-Aware Placement of Real-Time Virtual Machines in a Cloud Data Center

Cloud computing is an evolving paradigm which is becoming an adoptable technology for a variety of applications. However, cloud infrastructures must be able to fulfill application requirements before adopting cloud solutions. Cloud infrastructure providers communicate the characteristics of their services to their customers through Service Level Agreements (SLA). In order for a real-time application to be able to use cloud technology, cloud infrastructure providers have to be able to provide timing guarantees in the SLAs. In this paper, we present our ongoing work regarding a cloud solution in which periodic tasks are provided as a service in the Software as a Service (SaS) model. Tasks belonging to a certain application are mapped in a Virtual Machine (VM). We also study the problem of VM placement on a cloud infrastructure. We propose a placement mechanism which minimizes the energy consumption of the data center by consolidating VMs in a minimum number of servers while respecting the timing requirement of virtual machines.

Bandwidth adaptation in hierarchical scheduling using fuzzy controllers

In our previous work, we have introduced an adaptive hierarchical scheduling framework as a solution for composing dynamic real-time systems, i.e., systems where the CPU demand of their tasks are subjected to unknown and potentially drastic changes during run-time. The framework uses the PI controller which periodically adapts the system to the current load situation. The conventional PI controller despite simplicity and low CPU overhead, provides acceptable performance. However, increasing the pressure on the controller, e.g, with an application consisting of multiple tasks with drastically oscillating execution times, degrades the performance of the PI controller. Therefore, in this paper we modify the structure of our adaptive framework by replacing the PI controller with a fuzzy controller to achieve better performance. Furthermore, we conduct a simulation-based case study in which we compose dynamic tasks such as video decoder tasks with a set of static tasks into a single system, and we show that the new fuzzy controller outperforms our previous PI controller.

Exact and approximate supply bound function for multiprocessor periodic resource model: unsynchronized servers

The Multiprocessor Periodic Resource (MPR) model has been proposed for modeling compositional real-time guarantees of real-time systems which run on a shared multiprocessor hardware. In this paper we extend the MPR model such that the execution of virtual processors (servers) is not assumed to be synchronized i.e., the servers can have different phases. We believe that relaxing the server synchronization requirement provides greater deal of compatibility for implementing such a compositional method on various hardware platforms. We derive the resource supply bound function of the extended MPR model using an algorithm. Furthermore, we suggest an approach to calculate an approximate supply bound function with lower computational complexity for systems where calculating their supply bound function is computationally expensive.

Throughput Propagation in Constraint-Based Design Space Exploration for Mixed-Criticality Systems

When designing complex mixed-critical systems on multiprocessor platforms, a huge number of design alternatives has to be evaluated. Therefore, there is a need for tools which systematically find and analyze the ample alternatives and identify solutions that satisfy the design constraints. The recently proposed design space exploration (DSE) tool DeSyDe uses constraint programming (CP) to find implementations with performance guarantees for multiple applications with potentially mixed-critical design constraints on a shared platform. A key component of the DeSyDe tool is its throughput analysis component, called a throughput propagator in the context of CP. The throughput propagator guides the exploration by evaluating each design decision and is therefore executed excessively throughout the exploration. This paper presents two throughput propagators based on different analysis methods for DeSyDe. Their performance is evaluated in a range of experiments with six different application graphs, heterogeneous platform models and mixed-critical design constraints. The results suggest that the MCR throughput propagator is more efficient.

Adaptive multi-resource end-to-end reservations for component-based distributed real-time systems

Complexity in the real-time embedded software domain has been growing rapidly. The component-based software development approach facilitates the development process of such software systems by dividing a complex system into a number of simpler components. Resource reservation techniques have been widely used for providing resources to real-time software components. In this paper we target real-time components operating on a distributed resource infrastructure. Furthermore, we target a class of software components which demonstrate dynamic resource consumption behavior. A prime example of such components is a multimedia software component. In the paper, we present a framework supporting multi-resource endto- end resource reservations. We reserve resource bandwidths on both processor resources as well as on the network resources. The proposed framework utilizes a Multiple Input Multiple Output (MIMO) controller which adjusts the sizes of reservations tracking the dynamic resource demands of the software components. Finally, we present a case study using a multimedia component to demonstrate the performance and efficiency of our framework.

On Component-Based Software Development for Multiprocessor Real-Time Systems

Component-based software development provides a modular approach to develop complex software systems. In the context of real-time systems, it is desirable to abstract the timing properties of software components using an interface for each component. The timing properties of the whole system, composed of multiple components, is studied using the component interfaces. In this paper we focus on periodic interface models. In the case of components developed for single processor platforms, for examining the system schedulability, the interfaces can be regarded as periodic tasks. Thus, making it possible to use the conventional schedulability analyses for the system level schedulability test. In the case of components developed for multiprocessors, since interfaces may have utilization larger than 100% of a single processor, it is not possible to directly use the component interfaces for the system schedulability test. Therefore, the interfaces have to be decomposed before performing the system level schedulability test. In this paper, we target the special case of partitioned EDF for scheduling the components integrated on a multiprocessor. Therefore, the system level schedulability test is equivalent to finding a feasible allocation of component interfaces on the multiprocessor. We propose two algorithms for allocating the multiprocessor periodic interfaces. In addition, we propose an orthogonal approach for developing component-based real-time systems on multiprocessors in which components with utilization more than 100% of a single processor are divided into smaller subcomponents before abstracting their interfaces. We show, through extensive evaluations, that our alternative approach significantly reduces the interface overhead.

Adaptive hierarchical scheduling framework: Configuration and evaluation

We have introduced an adaptive hierarchical scheduling framework as a solution for composing dynamic realtime systems, i.e., systems where the CPU demand of its tasks are subjected to unknown and potentially drastic changes during runtime. The framework consists of a controller which periodically adapts the system to the current load situation. In this paper, we unveil and explore the detailed behavior and performance of such an adaptive framework. Specifically, we investigate the controller configurations enabling efficient control parameters which maximizes performance, and we evaluate the adaptive framework against a traditional static one. Furthermore, we demonstrate the results of our investigation using a practical multimedia case study in which we simulate the timing behavior of video decoding tasks running on our proposed framework. In addition, we compare the results of using our framework with the results of using static resource allocation approach.