Download cfghk - WOCC 2015

Embedded Computing and
Software Platform for Wireless
Networks
Prof. Zhaohui Wu, Ph.D
College of Computer Science
Zhejiang University, Hangzhou
310027, P.R.China
Outline
 Embedded Software in China
 Embedded Software Industry in China: Status and
Challenges
 Key R&D Projects about Embedded Software in China:
863 plan on software subject
 Embedded Computing: Status and Trend
 Sensor Networks are new Challenges
 Application Opportunity
 Technical Challenges for Embedded Computing
 Our Works on Embedded Software for
SmartPhone Computing and Sensor Network
Software Market in China (2004y)
 Package software(478亿元,
478,000,000,000)
 System Software（89亿元,89,000,000,000）
 OS（44.19亿元, 44190,000,000）
 Database Management System（20.06亿元,
20060,000,000）
 Software tools（25.71亿元, 25710,000,000）
 MiddleWare （6.98亿元, 6980,000,000）
 Application Software（304.53亿元,304,530,000,000）
 General Application Software（Security 、OA、MIS、
Game Software、CAI）
 Domain-Special Application Software（Telecom、
finance、government ）
 Embedded Software（673.62亿元, 673,620,000,000）
 Software Service and System Integration （937.48亿
元,937,480,000,000）
Embedded Software Market in China(1)
Embedded Software Market in China(2)
Embedded Software in China : Summary
The industry of embedded software is growing
dramatically and plays a more and more important
roles in software industry of China.
The biggest market of Embedded Software is
Telecom, including cell-phone, switch-class
and Router-class devices.
The three new domains for Embedded Software
in next 5 years in China: Smart-phone(ep. For
3G), IPTV, and e-auto
Embedded Software on mobile-phone
Embedded Software: various Hand-held Devices within
wireless connection for mobile computing
Tablet PC
Pocket PC
Desktop PC
Phone PC
Wrist PC
无线连接
Embedded Software on IPTV devices
Home-internet: new kind network for e-home IA
Digital Home Renderer
(DHR) Devices
Digital Home Server
(DHS) Devices
Transparent Connectivity
LCD TV
Digital Media Adapter
•
•
•
•
•
Wi-Fi / Wireless Protocol
Internet Protocol (IP)
Universal Plug and Play (UPnP)
Infrared Connectivity
Streaming Media
Video
Transmitter
Personal Video Player
PC
Multimedia Center
MP3 Player
DVD Player
Personal Video Recorder
Home Gateway
Media Storage
IP Camera
Wireless Video Router
Embedded Software on Auto-electron
Embedded Software: mesh network for e-Auto
汽车电子：占整车价值的45%；
围绕嵌入式LAN的改造是未来汽车电子发展的趋势
The market of Embedded Software for Medical Device is
also huge, wireless embedded network for Medical
Device will have a new chance
 Medical
 Xray system – Dicom, Imaging
 Biomedical Instrumentation
 Medical Infusion Pump
Embedded Software for Wearable Computing
will become more important in future: body
wireless network
Overview about 863 plan on
software subject
应用软件层
应用示范
开
源
社
区
测
试
体
系
技
术
标
准
中间件软件层
网络软件
运行环境
网络软件
开发环境
智
能
手
机
通
信
设
备
信
息
电
器
汽
车
电
子
面向领域的
嵌入式
软件平台
基础软件层
2017年5月25日星期四
数据库管理系统
办公软件
服务器操作系统
桌面操作系统
嵌入式
操作系统
15
The Goal and Tasks for 863 subject
on Embedded Software
 Real-time Embedded OS and Component Ware
 Application-Specific Embedded Software Platform
 For Telecom, for e-home and e-Auto
 Telecom-Oriented Embedded Software Platform
funded by 863 in the amount of
80,000,000RMByuan
 Cell-phone or smart-phone
 Switch-class Telecom Device
 Router-class Telecom Device
 Software Platform for new era network such as Ad
hoc and Sensor Network will be considered next 5
years
863手机嵌入式软件联盟
863手机嵌入式软件联盟CMSTA
 联盟宗旨
 嵌入式操作系统及软件平台的产业化, 形成面
向手机产业链的技术辐射型联盟组织CMSTA
（www.cmsta.org, 也称863手机软件技术联盟）；
 联盟模式
 独裁制: WinTel 产业联盟
 联邦制: OMG联盟
 联合国制: W3C, GGF
 组织形式
 联合国型：轮值主持
 技术辐射型联盟：技术辐射产业
 开放性的工作组：动态运作
 主要任务
 共同制定共性规范接口与数据交换格式，
 研制平台，实现技术共享与知识产权保护，
 推动联盟单位互补双赢合作
863手机嵌入式软件联盟的主要工作
 联盟工作
 2003年4月科技部领导提议，2003年8月四家单位发起成立
 目前联盟单位36个：
 手机厂商：12家（中兴通信、波导股份、普天东信、海尔通信、TCL移
动通信、联想移动、南方高科、海信通讯、大唐移动、华为科技、康
佳、科健）
 操作系统厂商：3家（凯思Hopen，科银京成Delta，SmartOs）
 应用软件厂商：16家（汉王科技、中科模识、茁壮网络等）；
 运营商: 2家(中国移动、联通)
 手机设计集成商：德信无线
 召开联盟全会及工作组会议21次
 制定了二个规范
 863手机软件联盟CMSTA的嵌入式基础平台软件参考模型与规范
接口（推行版 1.0.4）
 制定了手机软件联盟CMSTA的API符合性认证规范(A/B/C三级认
证)
Trend of Embedded Computing
 Embedded OS:
 More tiny and flexible, component ware,
 real-time, high-performance, high reliability, More
Security
 Power Aware Support: Just In time Power
 Development Environment IDE for Embedded
Software:
 Model-based Design
 Visualization
 Integration
Trend of Embedded Computing
 Embedded Software Platform
 Domain Specific Platform
 Open Standardization
Common Component Builder
 Open Source Driver
Java/Linux based Solutions
 Systematism Driver
Integration of OS, Embedded Middleware, and
Domain-Specific Platform
Trend of Embedded Computing
 New areas: new challenges
SOC(System on Chip): SoC and IP create new
opportunities for embedded software industry.
Internet integration: Deep-Embedded Network
Sensor Network and Ad hoc network: upsetkind network need a new genus of embedded
software
Ubiquitous Computing Environment and Smart
World
Sensor Network: Ongoing features and Goals
 Tight constraints
 Computation power
 Storage
 Power supply
 Traffic pattern
 Little activity in life period
 Tense traffic in short time
 Highly correlated traffic
 Environment
 Dynamic changing
Potential Applications
 Biomedicine
 Surgically implanted sensors monitoring a patient’s health
 Environmental monitoring
 Forest fire monitoring
 Local temperature gradients
 Spatially correlated temperature gradients
 Hazardous environment exploration
 Chemical concentrations
 Monitoring safety of children in neighborhood
 Military tracking, targeting
 Intrusion detection, identification (acoustic, imaging sensors)
 Spatially correlated troop, tank movement (seismic sensors)
 Monitoring foot paths: “Smart breadcrumbs”
 Inventory control
 Monitoring location of valuables in home
 Tracking inventory in ship-board containers
Technical Challenges for Embedded Software
 Tight resource constraints
 Energy
 Communications range, bandwidth
 Computation, storage (but not as constrained as energy and
communications – computation is often used to reduce
communication)
 Dynamically changing bandwidth, range, and computation power
 Dynamically changing network topology




Battery depletion
Node failure
Node mobility
Unreliable links (noise, jamming)
 Interactions
 Computation constraints lead to uneven power depletion which
leads to network topology changes
 Correlated bursts of traffic across neighboring nodes -- not a
collection of independent point-to-point flows – violating the design
assumptions of common media access protocols
Ongoing Projects in Our Lab
 SMART: Our Vision about Embedded
Computing and Systems
 SmartOS: Embedded Software
infrastructure for Smart-phone Computing
 SmartScud: A Context-Aware Lightweight
Embedded Software Platform for Sensor
Network
SMART：Our Vision
 SMART (Sub-coMponet ARchiteTure)
 Guidelines
 Embedded systems based on components of thingranularity
 flexible, optimized and customized
 Objectives
 high-quality codes
 high reliability
 real-time
 customized applications
 Application-specific OS
 SMARTOS: An CMSTA-compatible Embedded OS for
Smartphone
SMART Reference Model of
SmartPhone System
Application Software
Cross-platform IDE
Dynamically
Application
Abstract
Layer
G
U
I
Multimedia
Security
Embedded
DB
Java &
Internet
SMART-EOS
Resilient
Communication
Protocol
Stack
Process/Thre
ad
Management
Time
Management
File
Management
Power
Management
Device
Management
Data
Communication
Communication
Synchronization
& Exclusion
Management
Universally Adjustable Hardware Abstraction Layer
Embedded Operating System
 SMARTOS
Highly reliable &
component-based
embedded OS
ASOS (Application
Specific OS) Model
"Pipeline kernel" of
3rd-generation
micro-kernel
CMSTA-compatible
Cross-platform IDE
 Model-based Design Tools
 Model
 Visualization
 SMART-IDE
 Editor
 Linker
 Debugger
 Simulating environment
 Plug-in toolkit
 Memory leaking analyzing tools
 Coverage checking tools
 Performance measuring tools
 Static code checking tools
Embedded DBMS
 Concurrent control
 Real-time transaction commit protocol
 Data migration
 Online data update
 Database backup and recovery tools
 Memory-resident DBMS
 Flexibility design of DBMS
Embedded Multimedia
 Multimedia network communication protocol




stack component library
Multimedia network communication device
26 BREW cellular phone games
Embedded OS based audio, video, stream
media support
MP3
Cooperative Enterprises
 Bird(波导)
 UTStartcom（UT斯达康）
 Huawei technology(华为)
 ZTE（中兴）




Start(实达)
EASTCOM（东信）
Ningbo Samsung（宁波三星）
General Electric(美国通用
电气)
Putian（普天）
Mitsubishi（三菱）
Holly Group(华立)
Supor（苏泊尔）
DITEL （迪科）
SOYEA technology（数源科技）
Cellular Phones
DSP-based Video
phone
PDAs
IP TV: internet-based Digital TV
SmartScud: Lightweight Embedded Software
Platform for Sensor Network
 SmartScud
SmartOSEK: Tiny OS for Sensor Network
ScudWare: Lightweight Middleware for Sensor
Network
Routing Algorithms for Sensor Network
 Applications
Demo for Smart Space for Driver
Further Applications on Environment Sensing
and Monitoring
SmartOSEK: embedded OS
 Project abstract:
SmartOSEK Applications
SmartDesigner
OSEK-compliant Real-time Tiny
Operating System Platform
ECU
 The size of kernel <10k.
 The context switch time is less than
10us.
 Processors:
 AT89C51、 Intel 196、 FreeScale MPC555、
68HCS12、MSP430、Philips LPC2119、TI
TMS320C2407、ADI Blackfin 533 etc.
SmartOSEK IDE
 preemptive, non-preemptive
 group-based preemptive scheduling.
CAN
SmartEditor
 Project features
 Supporting three schedule schemes:
AMT
SmartOSEK OS
SmartOIL
SmartAnalyser
SmartOSEK
OS Kernel
BSP
SmartDebugger
CAN Bus Driver
SmartTracer
SmartOSEK
COM
ADC Driver
8051 MPC555 HCS12 ARM MSP430 DSP
SmartOSEK IDE
 SmartDesigner: the graphic modeling tool with
graphic interface, component operation, automatic
code generation and error detecting.
 SmartOIL: the graphic configuration tool with
configuration and parsing of OIL files, graphic
scalability and customization of the system kernel.
 SmartAnalyser: the graphic schedulability analysis
tool.
 SmartTracer: the graphic tool which traces the
running processes of the OS kernel and enables
users to observe the kernel status in a step-bystep manner.
The Context-based Lightweight Middleware
Architecture for Sensor Network
 Project: ScudWare
LightWeight Context-aware Reconfigurable
Middleware
 Project overview
The project is oriented to the heterogonous sensor
network computing environment in dynamic changing.
Scudware has the the properties of configurability,
reusability, security and is compliant to mini-CCM
specification. The research focuses on adaptive scheme
with non-monotone QoS requirements and restricted
resource and context management system.
Features of ScudWare
 Component management
 Service Construction and
Programming
Mechanism
 Contextual infrastructure
for Dynamically
Changing Resource
 Adaptive resource
configuration
Demo for Sensing SmartSpace of
Driver
– ZCCM-based design
– CAN Bus control devices
Routing Algorithms for Sensor
Networks(1)
 EDC[1]: An Event-Driven Clustering Routing
Algorithm for Sensor Network
 suitable for the event-driven static distributed
Sensor Networks
 ACHGT[2] :A Clustering Routing Algorithm Using
Game-theoretic Techniques
 Individual sensors of the algorithm are modeled as players to
make decisions about whether electing to become clusterhead nodes by using node energy payoff functions to weigh
their own costs and benefits.
 This strategy can keep sensor nodes of lower remainder
energy out of being used up quickly.
Routing Algorithms for Sensor
Networks(2)
 CRAM[3]: An Energy-Efficient Clustering
Routing Algorithm using Multi-CHs-One-Cluster
method for Sensor Networks
 selects Multi-CH nodes in every cluster together as one
CH node to perform data fusion and data transmission
to the BS so as to improve energy efficiency and
enhance data transmission reliability
 Since one CH node in a cluster may be unreliable and
cause energy losing in case of node failure or
communication error.
Thanks!
Technical Challenges
--Sensor Node Design
 Simplicity
 Cost
 Battery power
 battery replacement
 Deployment
 Resolution, accuracy, etc.
 Data sampling rates
 Clock sync
Technical Challenges
--Data communication and aggregation
 Wired vs. wireless communication
 Distances, bandwidth, multiple sensors
 Shared vs. dedicated channel access
 simple MAC protocol
 Routing through multi-hop networks
 Sensors are possibly stationary
 Sensors go into “sleep mode” to conserve energy
 Re-routing
Technical Challenges
--Sensor localization
 Each device has an factory-configured ID
 Network deployment is unplanned
 Location determination
 GPS-based
 Triangulation
 Etc.
Technical Challenges
-- Others
 Security (integrity, authentication,
confidentiality)
 Data compression
References
[1] Zheng Zeng-wei, Wu Zhao-hui, Lin Huai-zhong, “An EventDriven Clustering Routing Algorithm for Wireless Sensor
Networks”, 2004 IEEE/RSJ International Conference on
Intelligent Robots and Systems (IROS 2004), Sendai, Japan,
September 28 - October 2, 2004.
[2] Zheng Zeng-wei, Wu Zhao-hui, Lin Huai-zhong,
“CLUSTERING ROUTING ALGORITHM USING GAMETHEORETIC TECHNIQUES FOR WSNs”, IEEE International
Symposium on Circuits and Systems (ISCAS 2004), Vancouver,
Canada, May 23-26, 2004.
[3] Zheng Zengwei, Wu Zhaohui, Lin Huaizhong, Zheng Kougen,
“CRAM: an Energy Efficient Routing Algorithm for Wireless
Sensor Networks”, 19th International Symposium on Computer
and Information Sciences (ISCIS 2004), Kemer - Antalya, Turkey,
October 27-29, 2004.