Lecture on Evolution of Network Computing By Dr. GOPINATH GANAPATHY

1. Lecture on Evolution of Network Computing By Dr. GOPINATH GANAPATHY

2. Microprocessor Technology • Optical Networking Technology • Storage Technology • Wireless Technology • Sensor Technology • Global Internet Infrastructure • www and web services • Open-Source Movement IT Infrastructure Evolution

3. Year Transistors 1970 2,300 1975 5,000 1980 29,000 1985 275,000 1990 5,500,000 2000 42,000,000 2010 1,000,000,000 (One billion) 2.1 Microprocessor Technology Transistor Integration • Exponential growth transistors on a single chip from SSI, LSI, VLSI, ULSI, etc. (continued…)

4. 2.1 Microprocessor Technology…(continuation) • Computing Power • The Personal Desktop Computer of to-day processes One • billion floating point operations per second • The processing power of high-performance computing • systems processes has grown from 57.9 Gflops in 1993 to 40 • TFlops in 2003. • This is expected to be bumped up against the Laws of Physics • in 2017 through new technologies like Optical Lithiography. • The computing power will be taken on its next exponential • journey through future technologies like • Nuclear Magnetic Resonance • Ion Traps • Quantum Dots • Josephson Junctions • Optical Chips

5. 2.2 Optical Networking Technology • Developments in fiber optics and related technologies have increased internet traffic to many folds. • One of the key optical technologies, Dense Wave Division Multiplexing (DWDM) is responsible for this magic development. • In 1980s, system supported two widely separated wavelengths (1310nm and 1550nm) and each wavelength supported a traffic of a few megabits per second. • Over the years, the wavelengths increased from 2 to 8, reducing the spacing between wavelengths from 400 GHz to 12.5 GHz that increased the bit rate of each wavelength from a few Mbps to 10 Gbps. • A typical fiber system can now carry upto 640 Gbps • Use of MEMS (Micro Electro Mechanical Systems) and other technologies, replacing the costly electrical interfaces, and the tremendous capacity of bandwidth power in optical fibres, are cutting the price of bandwidth products aggressively downward.

6. 2.3Storage Technology • Over the last decade, disk storage capacity has • improved faster than Moore’s law of processing • power. • Current 2.5 inch hard disks can store around 30GB per disk platter • New disk head technology that enable hard disk drive • recording densities of upto 300 giga bits per square • inch, developed by Fujitsu Corporation, will increase • the storage capacity to 180 GB per platter in two to • four years time.

7. 2.4Wireless Technology • Wireless Technology is fast growing • Many countries employ 3G Technology that promises to • bring data at a rate of 1Mbps to a wireless device. • Fixed wireless Technologies such MMDS (Multipoint • Microwave Distribution System) and LMDS (Local • Multipoint Distribution System) in expected to revolutionize bandwidth access when the neat generation system an available. • IEEE is developing a family of standards (802.11,802.11a, • 802.11b, 802.11g) to govern wireless network

8. 2.5 Sensor Technology • Advancement in Sensor Technology has provided • the Sensor, the ability to collect information about • their surroundings and to pan the information about to • upstream system for analysis and processing through • various integrated communication channels. • Optical fibre Sensors have been in use since 1960s. • Wireless sensors with embedded RFID (Radio • Frequency Identification Devices) are being developed

9. 2.6 Global Internet Infrastructure • The internet started in the late 1960s when the US defense department’s Advanced Research Projects Agency developed ARPHA NET technology to linkup the networks of computers. • The network initially served academic researches, non-profit organizations and government agencies. • The National Science Foundation Network (NSFNET) gave US universities a nationwide high speed communications backbone for exchanging information. • During early 1990s NSFNET was privatized and became the Global Internet we know today (continued…)

10. 2.6 Global Internet Infrastructure... (continuation) • The first and the most significant change that has occurred over the last few years is a deep appreciation for open and standard’s, base systems. • The second key benefit is the substantial research, development, an implementation of security, security standards, an encryption for the public internet. • The third important development is the deployment of an impressive global networked infrastructure that is now ready to be harvested by new technologies such as GRID Computing.

11. 2.7 WWW and Web Servers • After 23 years of starting internet, the world wide web is • introduced to the public in 1993 • The World Wide Web turn out to be – “ killer app” that • lead to the expensive growth of the internet. • Today there are roughly 2 million hosts on the internet, • up from 1 million in 1993. • Web services provide a systematic and extensible frame • work for application to application interaction, built on • existing web protocols and based on open standards. (continued…)

12. 2.7 WWW and Web Servers ... (continuation) • The web services frame work is divided in to three areas • - Communication protocols, service description and • service discovery • SOAP - Simple Object Access Protocols enables communications among web services. • WSDL - Web Services Description Language provides a formal, computer readable description of web services • UDDI - Universal Description, Discovery and Integration Directory serves as a registry of Web Services Descriptions.

13. 2.8 Open Source Movement • The ubiquity of the internet and the world wide web has • also spawned a powerful movement per Open Source • software development. • Starting in 2001-2002, major vendors such as DELL, • HP, IBM, ORACLE and SUN announced in various • ways that they would begin supporting open source • products.

14. GRID Computing Technology - An Overview

15. In 1995 Super Computing conference in San Diego, 11 high speed • networks were used to connect 17 sites with high- end computing • resources for a demonstration to create one super “meta computer” • The project ‘GLOBUS’ lead by Ian Foster of ANL and Carl • Kesselman of University of Southern California created a suite of • tools that laid the foundation for Grid Computing. • In 1997, in Super Computing Conference, 80 sites worldwide • running software based on GLOBUS Toolkit were connected • together. • In 1997, Entropia was launched to harness the idle computer • worldwide to solve problems of scientific interest. History

16. A branch of computer science that concentrates on developing supercomputers and software to run on supercomputers. A main area of this discipline is developing parallel processing algorithms and software: programs that can be divided into little pieces so that each piece can be executed simultaneously by separate processors High performance computing

17. High performance, massively parallel computers built primarily out of commodity hardware components, running a free – software operating system such as Linux or Free BSD and interconnected by a private high speed network Connected to the outside world through only a single node AMD Athlon based cluster at University of Heidelberg in Germany was tested at 825 Gflops making it the 35th fastest high performance computer in the world Cluster Computing

18. Peer to Peer Computing • Decentralization • To overcome a performance bottleneck and • a single point of failure ofClient-server model • P2P Computing • – Napster, Gnutella, Freenet, JXTA • Napster • –Centralised model • Gnutella • – Decentralised model

19. Techniques learned in high-performance and cluster based distributed computing to utilize the vast processing cycles at users desktop. SETI@home, Parabon, Entropia In SETI Large Compute Intensive Projects are coded so that the tasks are broken down into smaller subtasks and distributed over the internet for processing. After completion of the task the central server aggregates the information and compiles the result. Internet computing

20. Grid Computing enables virtual organizations to share geographically distributed resources as they pursue common goals, assuming the absence of central location, central control, omniscience, and an existing trust relationship Virtual organization can be large or small, static or dynamic Resource is an entity to be shared that can be computational or storage resource Sensors and bandwidth are also used in the virtual organization In grid environment the resources do not have prior information about each other nor do they have pre- defined security relationships Grid Computing

21. Grid Implementation is trying to run program X using the resource at site Y subject to virtual community policy, p, providing access to data at Z according to the policy q. The Program X has to be able to work in an environment Y that could be heterogeneous and geographically dispersed. The Coordination of the use of the resources at sites Y and Z under the various restrictions and their usage as defined by policies p and q. Coordinated use of resources is more complicated when it is attempted across geographical and organizational boundaries. Grid Computing Model (continued…)

22. Grid Computing Model…(continuation) Key questions that come up in sharing resources across boundaries • Identify and authentication • Authorization and policy • Resource discovery • Resource characterization • Resource allocation • Resource Management • Accounting / Billing / Service level Agreement

23. Application Application Internet Protocol Architecture “Coordinating multiple resources”: ubiquitous infrastructure services, app-specific distributed services Collective “Sharing single resources”: negotiating access, controlling use Resource “Talking to things”: communication (Internet protocols) & security Connectivity Transport Internet “Controlling things locally”: Access to, & control of, resources Fabric Link Grid Computing Architecture Model

24. Applications Languages/Frameworks Collective Service APIs and SDKs Collective Service Protocols Collective Services Resource APIs and SDKs Resource Service Protocols Resource Services Connectivity APIs Connectivity Protocols Local Access APIs and Protocols Fabric Layer Grid Computing Architecture Model (in detail)

25. Fabric layer includes the protocols and interfaces that provide access to the resources that are being shared. • Connectivity layer defines core protocols required for grid specific network transaction. • – Utilizes internet protocol • – Utilizes core grid security protocol (grid security infrastructure) • Resource layer protocol required to initiate and control sharing of local resources • – Grid resource allocation management – remote allocation, reservation, monitoring and control of resources. • – Grid FTP – High performance data access and transport • – Grid resource information service - access to structure and state information Grid Computing Architecture Model (in detail) (continued…)

26. Collection layer provides system oriented capabilities for wide scale deployment – Includes index or meta-directory services for creating custom view of the resources – Resource brokers discover and allocate resources based on defied criteria. Application Layer defines services that are targeted towards a specific application or a class of applications. Grid Computing Architecture Model (in detail) (continuation)

27. GRID Protocols • Grid security Infrastructure • Grid Resource Allocation Management • Grid File Transfer Protocol • Grid Information Services

28. Security is defined in the resource layer of the grid architecture Resources located in different administrative domains makes it complex. Different requirements by users, resource owners and developers. – Users expect security system, easy to use, single sign-on capability, allow for delegation and support all key applications. – Resource owners expect local access control, have robust and detailed auditing and accounting and able to integrate with local security infrastructure. – Developers expect robust API/SDK to allow direct calls to various security functions. GRID Security Infrastructure (continued…)

29. Extensions for secure socket layer / transport layer security and X.509 to allow single sign-on and delegation. Generic Security Service provides functions for authentication, delegation and message protection. GRID Security Infrastructure... (continuation)

30. Allows program to be started on remote resources Resource specification language is a common notation for exchange of information between applications, resource brokers and local source managers. Types of Resource specification language – Resource requirements – machine type, number of nodes, memory, etc. – Job configuration – directory, executable, arguments, environment GRAM-2 protocol includes multiple resource types, web service protocols. GRID Resource Allocation Management Protocol

31. Grid File Transfer Protocol • Extension of standard FTP protocol with interoperability • with existing servers • Provides for striped / parallel data channels, partial files, • automatic and manual TCP butter size settings, • progress monitoring and extended restart functionality

32. Provides access to static and dynamic information regarding a grid’s various components. Types of grid information services – Grid resource information service – information about a specific resource – Grid index information service – aggregate directory service, provides a collection of information that has been gathered from multiple service. Grid Information Services

33. Reference implementation of the grid architecture and grid protocols. Open architecture, open source software toolkit. Tools and libraries. Security - Supports GSI Resource Management – Supports GRAM Data Managements – Supports Grid FTP Information Services- Supports GIS Globus Toolkit

34. Tries to marry web service architecture with grid computing architecture Similarity between grid technology and web services is the sharing of resources and facilitating the creation of virtual organization. The difference between grid technology and web service is that web service addresses persistent services while grid supports transient services. Grid service specification provides detailed specification for – How grid service are created and discovered – How grid service instances are named and referenced – Interfaces that define any grid service. Open Grid Service Architecture

35. Standard body governing the grid community. The mission is to focus on the promotional development of grid technologies and applications via the development and documentation of best practices, implementation guidelines and standards with an emphasis on rough consensus and running code. Formed by working groups and research groups Working group’s intention is to develop one or more specific documents aimed generally at providing specifications, guidelines or recommendations. Research group intends to explore an area where it may be premature to develop specification. GLOBAL GRID FORUM

36. TYPES OF GRIDS • Departmental grids • Enterprise grids • Extraprise grids • Global grids • Compute grids • Data grids • Utility grids

37. Deployed to solve the problems of a particular problem Cluster grids – term used by Sun Microsystems to provide a single point of access to users. Infra grids – term used by IBM to define a grid that optimizes the resources within an enterprise. Departmental Grids

38. Resources spread across an enterprise and provides service to all users. Runs behind the corporate fire-wall. Intra grids – used by IBM for sharing resource among different groups within an enterprise. Campus grids – used by Sun Microsystems to enable multiple projects or departments to share computing resources. Enterprise Grids

39. Establish between companies, their partners and their customers through a virtual private network. Extra grids – used by IBM to enable sharing of resources with external partners. Partner grids – grids between organization within similar industries to collaborate on projects. Extraprise Grids

40. Establish a public internet. Allow users to tap external resources. Inter grids – used by IBM to provide every ability to share compute and data / storage resources across the public web. Global Grids

41. Created solely for the purpose of providing access to computational resources. Desktop books – leverage computer resources of desktop computers. Server grids – grids limited to server resources High-performance / cluster grids – grids for super computers or HPC clusters Compute Grids

42. Grid deployment that require access to and processing of data Commercial compute resources that are maintained and managed by a service provider. Service grids – provide access to resources that can be purchased by corporations to augment their own resources Data Grids Utility Grids

43. Independent from the internet infrastructure Peering points serve as a point of connection between two or more networks. Star-light and Netherlight are established for handling research and academic grid networks GRID Networks

44. Perfect parallelism – divided into set of processes that require little or no communication, eg. Monte Carlo simulation. Data parallelism – operation performed on many data elements simultaneously. eg. Search different parts of a database. Functional parallelism – multiple operation performed simultaneously with each operation addressing the particular part of the problem. Eg. Power plant simulation. Granularity – the ratio of amount of computation done by an application to the amount of communication. GRID Application Characteristics

45. Application Integration [1] Message Passing Interface (MPI) and Parallel Virtual Machine(PVM)

46. Sleeper programs National security Philanthropic computing GRID Computing And Public Policy

47. Thank You