Recursion,Virtualization: basic concepts for Networking Yannis Corovesis (PhD Comp.Science) {ycor AT iit.demokritos.gr} Instititute of Informatics & Telecomms Summer School 2011 NCSR DEMOKRITOS Network problems Current tech. 2020 Address exhaustion Sizeglobal routing tables (in) Security Mobile Internet difficult Complex mechanisms for QoS Lack global names for applications Official reports Blumenthal, Clark 2001, Rethinking the design of Internet RFC 1287 T.Roscoe, The end of Internet Architecture HOTNETS 2005 IAB workshop 2007 http://tools.ietf.org/html/draft-iab-rawsreport-01 Olivier Martin, State of the Internet challenges ahead NEC 2007 http://www.ictconsulting.ch/presentations.ht ml Proposals for a new Network (s) McKeown, Girod Clean-Slate Design STANFORD Uni. 2006 http://cleanslate.stanford.edu/ Future Internet Assembly EU http://www.future-internet.eu/home/futureinternet-assembly.html Future Internet Architecture US http://www.nets-fia.net/ Personal experience Developing Internet in Greece http://ariadne-t.blogspot.com (in greek) CV like summary (gr, en) http://ycor.wordpress.com/ A 20+ question : What is Net ? Meta-artificial.blogspot.com Posts about network theory awareness: May 2009, Jan 2011, Jan 2012, Feb 2012 Pls send me comments: ycor AT ariadne-t DOT gr or twits: twitter.com/ycor Start of the thread In 1990 I named the NRN project “net what ?” http://www.ariadne-t.gr and declare its existence to Stanford Research Institute that run NIC http://www.tinyurl.com/internetGR1992 In 1996 after some progress with basics I presented them to a biz conference on behalf Greek PTT http://www.tinyurl.com/internetOTE Layer-Packet-Peering-Client-Server Understanding Internet processes running on Node computers arranged in LAYERS over PACKET SWITCHING with PEERING protocols and CLIENT-SERVER programs Theory and Network This presentation considers there has been lots of engineering but little science … borrows the analogy from computing http://www.cs.kent.ac.uk/people/staff/dat/miranda/ctfp. pdf http://lambda-the-ultimate.org/node/2204 that theory helped fix computer architectural concepts and achieve convergence in order for us today to consider an ΙΝΤΕL code program to be the same entity to an AMD program and a Java program to be the same as a C program ! There is a need for something similar to occur in Networking. Metasystem Transitions Strowger Switch: automates a manually operated system http://www.seg.co.uk/telecomm/automat1.ht m ARPANET: enables hooking up of different vendor terminals http://en.wikipedia.org/wiki/ARPANET Evolution path Το 1974 NCP basic network protocol is replaced NCP by TCPIP to enable connection of different tech. networks http://el.wikipedia.org/wiki/TCP/IP A period of small and big network advances take place that fit the pattern “automation of manual operations” Automation Steps Routing EGP, BGP1, BGP2, BGP3, BGP4 Naming Domain Name System 1987 Introduce multimedia to text Internet ΜΙΜΕ 1991 http://en.wikipedia.org/wiki/MIME Network Information Resources e.g system ARCHIE http://en.wikipedia.org/wiki/Archie_search_engi ne Global web WWW what is automation? Abstractions needed J.POSTEL proposed to R.KHARE a study over tech documents at ISI with the goal to obtain a “picture” (1998) – reference R.Khare in Postel's orbituary F.BAKER (IAB/IETF) proposes PSEUDOAREA in an IETF email. A typewriter inspired A.TURING - book “the enigma” Α.Hodges Inspiration from Metasystem transition Valentin F. Turchin Recursion – Virtualization give a abstractions Network in stage Ν is a composed from An Automation-Quantum embedded upon Net in stage Ν-1 gradually spreading Auto-Q is a distributed application “running” on a virtual CPU, using a virtual MEMORY component (sumtotal of real Ms) and is using as a virtual BUS the existing Net to propage itself as a transformation Meeting RINA In 2009 googling (Net,Arch, Evolution, Patterns) I came across the book “Patterns in Networking Architecture” byJohn Day, the author says he has been searching the question "What is Network" for 35 years, initially as a graduate student building ΑRPANET, later writing RFCs, specialized in Protocol Formal Specification, in brief a Nework pioneer where sometime in 1994 he located “patterns” related to network fundamentals. He was an OSI model's specification reporter from 1990/4 he understands OSI better than me that in 1994 declated “OSI can be harmful” that missed its positive contribution. http://www.islab.demokritos.gr/gr/html/parousiaseis/im erida_1.htm Recursive InterNetwork Architecture John Day says we must first see the "what" and then proceed to "how " . If we do so, then TCP/IP splits the layers differently (horizontal split instead of vertical)), meaning TCP and IP should ne on the same layer mechanism but using diffent POLICIES. Left side is network Α and right side is network Β that have different scopes. If a an application process calls for connection between A and B then a new network C (LAYER) must be “stacked” over them. Network C is already there or created on-demand. RINA Ενα είναι το LAYER ! αλλά παίζει RECURSION κάνοντας ενθυλάκωση του εαυτού του μέσα σε ένα ίδιο LAYER οπότε απλοποιείται ο μηχανισμός πέφτει το κόστος και έχουμε φορητότητα, ασφάλεια, αέναη κλιμάκωση http://csr.bu.edu/rina http://pouzin.pnanetworks.com/documentspublicat ions.html Στις βασικές έννοιες ΑΝΑΔΡΟΜΗΕΙΚΟΝΟΠΟΙΗΣΗ στέκεται και ο ερευνητής Joe Touch από το ISI δημοσιεύοντας το 2011 μαζί με την ομάδα RINA του J.Day το μοντέλο DRUID http://www.isi.edu/touch/pubs/comnet2011.pdf The ΝΕΤ8 model Functional Prog. Paradigm inspiration (abstaction power, evolution logic) The NET is the COMPUTER ™ Meta-systems (as reference model) Αbstractions: Automation-Quantum, Recursion, Virtual Von-Neumann Machine ΝΕΤ8-RNA-RINA-DRUID In http://meta-artificial.blogspot.com“ I describe how I discovered and what it means for me “John's apocalypse ”. I notice that EU reseach shows somehow the same interest as I do: https://tnc2011.terena.org/core/presentation /63http://www.futureinternet.eu/home/future-internetassembly/budapest-may-2011.html