# 电子工程代写|嵌入式网络系统代写Embedded Networked Systems代考|ELEC5514

## 电子工程代写|嵌入式网络系统代写Embedded Networked Systems代考|Formalizing the Reconfiguration Functionality

The runtime reconfiguration process can be considered as a continuation of the system design process that is performed at runtime. As such it remains a domain knowledge intensive process. Design decisions-such as algorithm selection, parameterization, interconnection topology, task allocation-should be made on various levels based on the actual circumstances, and guided by design expertise, general systems engineering knowledge, etc. Also when dealing with this complexity aspect, the architectural separation of the primary and the reconfiguration functionalities (with clear “actuation interfaces” between them, see Fig. 2.4) results in designs with cleanly assigned responsibilities and manageable non-functional properties.

The implementation of the reconfiguration functionality can take various forms. A frequent implementation pattern is to develop a custom algorithm addressing the particular case in hand. In this case, the “design knowledge” is hard-coded into the implementation. Such approach is frequently chosen for its runtime performance, but-due to the eventual complexity of the reconfiguration challenge-the solution can be error prone. In addition, extending this built in “design knowledge” can be an overly demanding work because it may require thorough rewrite and extension of the existing custom code. An more efficient and economic alternative is to follow the “knowledge based approach” pattern, in which the “knowledge” and the “use of the knowledge” are clearly separated entities.

Figure $2.7$ shows a reconfiguration solution inspired by “knowledge based” pattern. As the name suggests the “knowledge based” pattern explicitly represents and uses knowledge. By knowledge, we mean any formal representation information, which is relevant in the context of making decisions about the using system resources to achieve the pre-set operational goals. In most of the approaches, the knowledge is captured into models that describe certain aspects of the operation of the system. Knowledge elements also describe the current state of the system relevant to reconfiguration: this part is called dynamic knowledge.

## 电子工程代写|嵌入式网络系统代写Embedded Networked Systems代考|Task Models for Runtime Reconfiguration

Task models is a design pattern for reconfiguration extensively used in the DEMANES project. The idea behind task models is that the behavior of a (distributed) embedded system should be formalized as a set of interacting tasks. Consequently, the processing scheme implementing the runtime reconfiguration should be mapped into a task model. Each reconfiguration scheme define roles for tasks and fixes where they should be located. Here we consider only the most commonly used configurations, namely:

• Local monitoring, local reasoning, local actuation (LLL type reconfiguration)
• Full scale monitoring, local reasoning, full scale actuation (FLF type reconfiguration)
• Constrained monitoring, local reasoning, local actuation (CLL type reconfiguration)
• Constrained monitoring, constrained reasoning, local actuation (CCL type reconfiguration)
The interpretation of the terms is as follows:
• Local: the scope of the activity is restricted to the node hosting the PF targeted by the reconfiguration (the PF mentioned here may only be a part of the complete PF)
• Constrained: the scope of the activity is restricted to the a subset of nodes hosting the PF
• Full: the scope of the activity covers the total system (i.e. all nodes are involved)
The LLL type of reconfiguration is the simplest scheme: every node monitors its own execution state, reasons locally about local goals and the reconfiguration actions are restricted to the node itself. The FLF type of reconfiguration corresponds the centralized implementation of the reconfiguration: one assigned node collects all execution state information from all nodes comprising the system, carries out the reasoning (locally) and actuates components on all nodes. The CLL differs from LLL that the local reasoning uses information about the execution state of a subset of nodes, typically neighboring nodes. The CCL scheme relies on cooperative reasoning mechanisms: beside sharing execution states the reasoners cooperate during the reasoning process to establish consensus, i.e. they attempt to achieve system-wide optimality instead of selfish local optimality.

## 电子工程代写|嵌入式网络系统代写Embedded Networked Systems代考|Task Models for Runtime Reconfiguration

• 本地监控、本地推理、本地驱动（LLL 类型重新配置）
• 满量程监控、局部推理、满量程驱动（FLF 型重新配置）
• 受限监控、局部推理、局部驱动（CLL 类型重新配置）
• 约束监控、约束推理、局部驱动（CCL类型重新配置）
术语解释如下：
• 本地：活动范围仅限于承载重配置目标PF的节点（这里提到的PF可能只是完整PF的一部分）
• 受限：活动范围仅限于托管 PF 的节点子集
• 全：活动范围覆盖整个系统（即涉及所有节点）
LLL 类型的重新配置是最简单的方案：每个节点监控自己的执行状态、本地目标的本地原因以及重新配置操作仅限于节点本身。FLF 类型的重新配置对应于重新配置的集中实现：一个分配的节点从组成系统的所有节点收集所有执行状态信息，执行推理（本地）并启动所有节点上的组件。CLL 与 LLL 的不同之处在于，本地推理使用有关节点子集（通常是相邻节点）的执行状态的信息。CCL 方案依赖于合作推理机制：除了共享执行状态之外，推理者在推理过程中合作建立共识，即

myassignments-help数学代考价格说明

1、客户需提供物理代考的网址，相关账户，以及课程名称，Textbook等相关资料~客服会根据作业数量和持续时间给您定价~使收费透明，让您清楚的知道您的钱花在什么地方。

2、数学代写一般每篇报价约为600—1000rmb，费用根据持续时间、周作业量、成绩要求有所浮动(持续时间越长约便宜、周作业量越多约贵、成绩要求越高越贵)，报价后价格觉得合适，可以先付一周的款，我们帮你试做，满意后再继续，遇到Fail全额退款。

3、myassignments-help公司所有MATH作业代写服务支持付半款，全款，周付款，周付款一方面方便大家查阅自己的分数，一方面也方便大家资金周转，注意:每周固定周一时先预付下周的定金，不付定金不予继续做。物理代写一次性付清打9.5折。

Math作业代写、数学代写常见问题

myassignments-help擅长领域包含但不是全部: