Advanced Operating Systems Lecture notes Dr Clifford Neuman

Advanced Operating Systems Lecture notes Dr. Clifford Neuman University of Southern California Information Sciences Institute Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Announcements v. New course in Spring - Trusted Computing q http: //ccss. usc. edu/599 tc q Friday’s at 1 PM v. Paper is due today q If not postmarked today, but finished by Friday the 8 th, a small penalty will apply. v. Final exam q Friday December 8 th at 2: 00 PM q Location is ZHS 252 & ZHS 352 Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

CSci 555: Advanced Operating Systems Lecture 14 – December 1, 2006 Special Topics, Review Dr. Clifford Neuman University of Southern California Information Sciences Institute Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Requested Topics v Grid Computing & Data Grids v OS Trends - Is the OS still relevant v Embedded Systems v Hardware Abstraction v What happened to the research systems and people v Future of Distributed Systems v Content Delivery Networks v Simulating systems v Peer to Peer v Windows Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Grids v Computational grids apply many distributed system techniques to meta computing (parallel applications running on large numbers of nodes across significant distances). q Libraries provide a common base for managing such systems. q Some consider grids different, but in my view the differences are not major, just the applications are. v Data grids extend the grid “term” into other classes of computing. q Issues for data grids are massive storage, indexing, and retrieval. q It is a file system, indexing, and ontological problem. Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Is the OS still relevant v What is the role of an OS in the internet q Are today’s computers appliances for accessing the web? v OS Manages local resources q Provides protection between applications q Though the role seems diminished, it is actually increasing in importance Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Embedded Systems v Process control / SCADA v Real time often a factor v Protection from external influences q i. e. dedicated bandwidth q Avoid general purpose interfaces v Are newer embedded systems really embedded? Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Hardware Abstraction v Many operating systems are designed today to run on heterogeneous hardware v Hardware abstraction layer often part of the internal design of the OS. q Small set of functions q Called by main OS code v Usually limited to some similarity in hardware, or the abstraction code becomes more complex and affects performance. Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Where are they Now v AFS, NFS, Athena, Andrew RPC, Andrew File System, Kerberos, Locus, HCS v Birrel, Lampson, Needham, Schroeder, Popek, Spector, Gifford, Saltzer, Jefferson, Cheriton, Mullender. Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Future of Distributed Systems v More embedded systems (becoming less “embedded”). v Stronger management of data flows across applications. v Better resource management across organizational domains. v Multiple views of available resouces. Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Content Delivery v Pre-staging of content v Techniques needed to redirect to local copy. v Ideally need ways to avoid central bottleneck. v Use of URN’s can help, but needs underlying changes to browsers. q For dedicated apps, easier to deploy Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Simulations v Need techniques to test approaches before system is built. q Simulations q Need real data sets to model assumptions. v Need techniques to test scalability before system is deployed. q Deployment harder than implementation q Emulations and simulations beneficial v Issues in emulation and simulation Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Peer to Peer v Peer to peer systems are client server systems where the client is also a server. v The important issues in peer to peer systems are really: q Trust – one has less trust in servers q Unreliability – Nodes can drop out at will. q Management – need to avoid central control (a factor caused by unreliability) v Ad hoc network related to peer Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Windows v XP, Win 2 K and successors based loosely on Mach Kernel. v Techniques drawn from many other research systems. v Backwards compatibility has been an issue affecting some aspects of it architecture. v Despite common criticism, the current versions make a pretty good system for general computing needs. Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE

Review for Final Design Problem v You have been hired by a new internet startup "Indelible" to build a peer-topeer backup system. You were hired very early in the design process, and as such you have not been provided with much guidance about what is required from such a system. Your first task is to identify the issues in building such a system, including 1) the choice of peers to which one backs up data, and 2) the list of peers from which one accepts backups, 3) how one identifies and organizes backed up objects, 4) how one protects such objects, and 5) how one decides which objects to backup. Performance, storage capacity, reliability, privacy, and fairness will be critical in the system you build. Discuss each of the five numbered issues above, mentioning the implications for the critical characteristics mentioned above. Pick one of the above issues and describe a design that exhibits favorable characteristics. Your score for these two parts will depend both on your solution (7 b 10 points) and on (7 c 5 points) the challenge of your choice of characteristics (i. e. , if you pick the trivial issues and characteristics you will get a poor score for challenge). Copyright © 1995 -2006 Clifford Neuman and Dongho Kim - UNIVERSITY OF SOUTHERN CALIFORNIA - INFORMATION SCIENCES INSTITUTE