Chapter 2: Transparent Computing and Computing Lecture 2 透 明 计 算 与 云 计 算 的 关 联 Prof. Zixue Cheng 程 子 学 University of Aizu, 会 津 大 学 Visiting Professor of CSU 1 Contents of the lecture Definition, Architecture Saas, PaaS, and IaaS Examples of Computing Relation between Transparent Computing, IoT, and Computing Comparison between TC and CC Relation between TC, IoT, and TC 2 1
What is Internet? : IT resources Scalable and Measurable Remotely access What is IT resources Physical/virtual servers Software Programs Services Storage devices Network Devices On Premise based and On based Consumers and providers 3 Scale up Processing power increase Storage increase Scale out Processing unit increase Storage unit increase Scale up and scale out Scale up (Vertical scaling ) Scale out (Horizontal scaling ) Merits Reliable Instantly available Automatic scaling No limitation Demerits Limited by HW capacity Expensive Additional resources are necessary 4 2
Computing Image (Concept) Scale out resources SaaS, PaaS, IaaS Software as a Service Mail, Game Virtual Desk Top Platform as a Service Runtime support DB Develop tools Infrastructure as a Service Virtual servers, Networks Virtualization abstraction of the physical recourses Flexibly use the physical recourses 5 Business motivation Motivation of the Proportional costs : Reduce the investment : small investigation can have IT infra. Increased scalability : dynamically depending on needs. Increased availability and reliability : Risk Increased vulnerability Reduced operational governance control Limited portability between cloud providers Multi Regional Compliance and Legal issues 6 3
SaaS Architecture of Computing (Fig. 4.16 textbook) Service 1 Service 2 Service 2 PaaS Environment 1 IaaS Virtual Server 1 Physical Server 2 Virtual Server 2 7 Comparison: Role and limitations Delivery Model SaaS PaaS IaaS Consumer Activities Uses and configures cloud service Develop, tests, deploys, and manages cloud services and based solutions Sets up and configures bare infrastructure and installs, manages, and monitors any needed software Provider Activities Implements, manages, and maintains cloud service Monitors usage by cloud consumers Pre configures platform and provisions underlying infrastructure middleware, and other needed IT resource, as necessary Monitors usage by cloud consumers, Provisions and manages the physical processing, storage, networking and hosting required Monitors usage by cloud consumers 8 4
Technologies for Computing (1) Data Center Data center Data center Rack Cluster 9 Technologies for Computing (1) Data Center Specialized IT infrastructure that houses centralized IT resources, such as servers, databases, and software systems Standardization and modularity Automation : Remote operation and management High Availability Security aware design, operation, and management Facilities Computing hardware Storage hardware, Network hardware 10 5
Technologies for Computing (2) Virtualization VM Guest OS and app. VM Guest OS and app. VM Guest OS and app. VM Guest OS and app. Virtual Machine Management (Virtual Machine Monitor: Hypervisor) Operating System (Host OS) Hardware 11 Crucial Differences between TC and CC (by Prof. Jingde Cheng, Saitama University) TC The TC is based on a deep analysis about the problems and their causes of the traditional computing systems from a viewpoint of end users. The TC is an original creation of new ideas. The TC has a clear definition to distinguish from others. The TC is technology oriented, user oriented. The TC is not so easy to understand. CC The CC is a natural evolution of existing techniques (distributed computing, grid computing, utility computing, SaaS, ). The CC is just a new name putting something together. There is no clear definition for the CC until now, although there are a lot of definitions, i.e., a cloud of definitions. The CC is business oriented, provider oriented. The CC is very easy to understand, and also very hard to understand. 12 6
Comparison of TC and CC (2) Motivation TC Because the tight coupling of software and hardware in traditional computing systems, current computer systems are not suitable for pervasive computing. Problems in traditional computing systems: Complexity, High total cost of ownership, Weak security, Not user friendly. The problems are motivated from a viewpoint of end users. CC Service providers try to provide simplified software installation, maintenance, and a centralized control over the software used. The problems are motivated from a viewpoint of service providers. 13 Comparison of TC and CC (3) Fundamental Ideas TC To separate hardware and software. To separate computation and storage. To provide end users with computation results/services but make the process of computation transparent to the users. CC : a pool of abstracted, virtualized, dynamically scalable, managed computing power, storage, platforms, and services are delivered on demand to external customers over the Internet. It is combining the cloud with computing that causes a lot of confusion! 14 7
Characteristics / Features Comparison of TC and CC (4) Characteristics/Features: TC vs CC Transparent Computing Computing User friendliness The purpose Better than traditional paradigm? Transparency The purpose A natural result of Virtualization requirement Virtualization A natural result of transparency requirement The purpose Dynamic scalability Bottleneck? The purpose Service model All as a Service (AaaS) SaaS, PaaS, IaaS Centralized servers Centralized servers 15 (What is NOT IoT?) What is IoT? US IoT Ubiquitous System = IoT? No, US IoT Cyber Physical System = IoT? CPS IoT Computing = IoT? IoT Internet = IoT? Internet IoT 16 8
Sensor layer: RF ID, camera, Actuator, robots Situation aware 物 联 网 的 结 构 体 系 IoT architecture Internet Big Data Things Network Layer TCP/IP, IPv6 6LowPAN, IP over things, conflict free Sensors Actuator Comm. and Coop. between Things, M2M, M2P, P2M, 17 An Example of IoT system, (IBM) River pollution monitoring http://www 06.ibm.com/jp/press/2010/11/2501.html Ask citizens using camera of iphone, to take and upload a picture of the river, flowing through the city, with answers to some simple questions: Water level, Speed, Garbage amount 18 9
Features of IoT (1) Specially Basically Sensing function Communication Smart App. & Services Scalability Openness Aware/Finding Intelligence/Smartness Real Time and Just in Time Conflict free/resolution Safety 19 Hot or Cool Features of IoT (2) More about Conflicts and Safety Requirement Resources Monitors with cameras in a smart city Traffic control Fireman, Police Service provider Time Sensors Actuators, Physical Objects Equipment Persons 20 10
Features of IoT (3) Safety Internet Poison Gas Temperature/ Pressure 21 Basic info 973 project 1 Beijing University of Posts and Telecommunication Leader: Huadong Ma, Totally 30 researchers, 2010.9 2015.9 Name: Research on basic architecture of IoT Associated departments 1. Tsinghua University 2. Shanghai Jiao Tong University 3. Institute of Computing Technology Chinese Academy of Science 4. Beijing University of Technology 5. Beijing University 6. Zhejiang A&F University 7. Xi an Jiao Tong University Base Theory Features Testbed Result Network information Theory Uncertainty Large scale heterogeneous network, uncertain information, Self adaptive services in dynamic environment Environment monitoring SCI 50, IEEE/ACM Trans. 20 Patents: 18, Standard: 2 PhD 30 Conference arrangement 2 3 Test bed 1 973 project 2 Tong Ji University(Maco) Leader: Wei Zhao Totally 32 researchers 2011.1 2013.8 Name: Research on basic theory and design methods of IoT 973 project 3 Shanghai Institute of Micro system and Information Technology Leader: Haitao Liu Totally 30 researchers 2011.1 2013.8 Name: Research on basic theory and application of IoT 1. East China Normal University 2. Institute of Computing Technology Chinese Academy of Science 3. Tsinghua University 1. The Institute of Software, Chinese Academy of Sciences 2. Beijing University of Posts and Telecommunications 3. Tsinghua University 4. East China Normal University 5. University of Science and Technology of China 6. Southeast University 1. Network calculus, 2. Dioid algebra, 3. Real time network scheduling and queuing theory 1. Correlation Model, Network Architecture; 2. heterogeneous networks convergence; 3. Hybrid information fusion. Information fidelity execution loyalty, Uncertainty information, Unpredictable information, Fuzzy information Strong correlation between the physical space and information space High mixed system, e.g. various devices, information; Information uncertainty. Smart grid Papers: 160 (Conf. 50) Books: 2 Patents: 5 Software copyrights: 2 PhD and Post Doc: 35 Master: 30 1. Smart grid; 2. Smart home; 3. Electric car Papers: 40 50, Patents: 15 Monograph: 1 National Standard: 2 4 International Standard: 1 Graduate Students:50 22 11
云 计 算, 透 明 计 算, 和 物 联 网 Comp., Transparent Comp., and IoT Comp. Scale out resources SaaS: Software as a Service Virtualization Transparent Comp TransOS as simple functions as possible OS and applications can be downloaded from servers Piece by piece in a flow manner ThansOS Other part Preserve Kernel of TransOS EnabledBIOS Hardware Terminal Internet Download piece by piece in a flow Network ThansOS Other part OS of Server BIOS Hardware Servers 23 Summary Basic of Computing Architecture Saas, PaaS, and IaaS Data center, virtualization Relation between Transparent Computing and Computing IoT, computing, and Transparent computing Understanding those computing paradigms and employing them at properly Comp. IoT Trans. Comp. 24 12