چارچوب مدیریت منابع چند لایه برای مدیریت منابع پویا در سیستم های سرمایه گذاری DRE

Enterprise distributed real-time and embedded (DRE) systems can benefit from dynamic management of computing and networking resources to optimize and reconfigure system resources at runtime in response to changing mission needs and/or other situations, such as failures or system overload. This paper provides two contributions to the study of dynamic resource management (DRM) for enterprise DRE systems. First, we describe a standards-based multi-layered resource management (ARMS MLRM) architecture that provides DRM capabilities to enterprise DRE systems. Second, we show the results of experiments evaluating our ARMS MLRM architecture in the context of a representative enterprise DRE system for shipboard computing.

Enterprise distributed real-time and embedded (DRE) systems, such as shipboard computing environments (Schmidt et al., 2001), airborne command and control systems (Loyall et al., 2001), and intelligence, surveillance and reconnaissance systems (Sharma et al., 2004), are growing in complexity and importance as more computing devices are networked together to help automate tasks previously done by human operators. These types of systems are characterized by stringent quality-of-service (QoS) requirements, such as low latency and jitter, expected in real-time and embedded systems, as well as high throughput, scalability, and reliability expected in enterprise distributed systems. Enterprise DRE systems have a range of QoS requirements that may vary at runtime due to planned (Li and Nahrstedt, 1999) and unplanned (Abdelzaher et al., 2003) events. Examples of planned events include mission goal changes due to refined intelligence and planned task-completion exceeding mission parameters. Likewise, examples of unplanned events might include system runtime performance changes due to loss of resources, transient overload, and/or changes in algorithmic parameters (such as modifying an air threat tracking subsystem to have better coverage). Dynamic resource management (DRM) ( Welch et al., 1998 and Hansen et al., 2001) is a promising paradigm for supporting different types of applications running in enterprise DRE system environments—as well as to optimize and reconfigure the resources available in the system to meet the changing needs of applications at runtime. The primary goal of DRM is to ensure that enterprise DRE systems can adapt dependably in response to dynamically changing conditions (e.g., evolving multi-mission priorities) to ensure that computing and networking resources are best aligned to meet critical mission requirements. A key assumption in DRM technologies is that the levels of service in one dimension can be coordinated with and/or traded off against the levels of service in other dimensions to meet mission needs, e.g., the security and dependability of message transmission may need to be traded off against latency and predictability. This paper describes a multi-layer resource management (ARMS MLRM) architecture we developed to demonstrate DRM capabilities in a shipboard computing environment. This environment consists of a grid of computers that manage many aspects of a ship’s power, navigation, command and control, and tactical operations (Schmidt et al., 2001) using standards-based DRM services that support multiple QoS requirements, such as survivability, predictability, security, and efficient resource utilization. Our ARMS MLRM was developed for the DARPA’s Adaptive and Reflective Middleware Systems (ARMS) program (dtsn.darpa.mil/ixodarpatech/ixo_FeatureDetail.asp?id = 6), which is applying DRM technologies to coordinate a computing grid that manages and automates many aspects of shipboard computing. We describe and empirically evaluate how the ARMS MLRM manages computing resources dynamically and ensures proper execution of missions in response to mission mode changes and/or resource load changes and failures, as well as capability upgrades.

This paper described a standards-based, multi-layered resource management (ARMS MLRM) architecture developed in the DARPA ARMS program to support dynamic resource management (DRM) in enterprise distributed real-time and embedded (DRE) systems. The ARMS MLRM is designed to enable enterprise DRE systems to adapt to dynamically changing conditions (e.g., during a tactical engagement) for the purpose of always utilizing the available computers and networks to the highest degree possible in support of mission needs under various operating conditions. The lessons learned while developing ARMS MLRM and applying it to a shipboard computing environment include: • ARMS MLRM research and experiments shows that dynamic resource management (DRM)—using standards-based middleware technologies—is not only feasible, but also can (1) handle dynamic resource allocations across a wide array of configurations and capacities, (2) provide continuous availability for critical functionalities—even in the presence of node and pool failures—through reconfiguration and redeployment and (3) provide QoS for critical operational strings even in the conditions of overload and resource constrained environments. • Enterprise DRE systems that are provisioned statically require a great deal of manual engineering effort to create and validate any (re)configuration and (re)allocation. This tedious and error-prone manual process marginalizes any advantage that might be gained by leveraging allocation and configuration as a means of application resilency. The DRM capabilities provided by ARMS MLRM help alleviate much of this inflexibility by consistently achieving well-balanced allocation under varying conditions, and permiting a range of automated, dynamic management of resources that greatly extend operational flexibility of the system without human intervention. Moreover, our experiments clearly demonstrate that the performance of ARMS MLRM is enhanced when DRM services are enabled—and in fact allow ARMS MLRM to operate in the presence of failures that cannot be accommodated by static resource management. • Enterprise DRE systems have a significant number of software architecture components that require QoS, configuration, and deployment information. Although ARMS MLRM provides an effective software infrastructure for DRM, it is a complex task to capture the required application information for all components. To deal with this complexity, we are exploring model-driven development (MDD) tools (Balasubramanian et al., 2005) that enable a software infrastructure consisting of standards-based middleware and components so that ARMS MLRM functions and workflow can be performed more effectively and productively.