Il Software Libero nella ricerca universitaria: l esperienza del gruppo di ricerca COMICS

Transcript

1 Il Software Libero nella ricerca universitaria: l esperienza del gruppo di ricerca COMICS /LQX['D\ Napoli, 25 Ottobre 2008 Antonio Pescapè, pescape@unina.it Dipartimento di Informatica e Sistemistica Università degli Studi di Napoli Federico II

2 Agend a ¾ Ringraziamenti ¾ Cosa significa fare ricerca ¾ Un settore a me caro: le reti ¾ La ricerca del gruppo COMICS ¾ Progetti OS di COMICS ¾ Conclusioni

3 La Ricerca ed il Metodo Scientifico (1/ 3) ¾ Il PHWRGRVFLHQWLILFR è la modalità tipica con cui la scienza procede per raggiungere una conoscenza della realtà RJJHWWLYD, DIILGDELOH, YHULILFDELOH e FRQGLYLVLELOH. ¾ Esso consiste, da una parte, nella raccolta di evidenza empirica e misurabile attraverso l'osservazione e l'esperimento; dall'altra, nella formulazione di ipotesi e teorie da sottoporre nuovamente al vaglio dell'esperimento. Source: wikipedia

4 La Ricerca ed il Metodo Scientifico (2/ 3) ¾ Quale ricerca: di base, applicata, industriale. ¾ Condivisione della conoscenza e dei risultati. ¾ Il sistema di referaggio e validazione:» 3HHU 5HYLHZ» &URVV&KHFNLQJ

5 La Ricerca e Internet

6 Da Inform atica Centralizzata a Inform atica Distribu ita ¾,QIRUPDWLFD &HQWUDOL]]DWD» Un elaboratore centrale, con molti terminali non intelligenti (terminali e stampanti)» *HVWLRQH FHQWUDOL]]DWD dell IT ( CED )» Un solo computer ( mainframe ), molti utenti» Grandi capacità di calcolo» Architetture multiprocesso e 6LVWHPL 3URSULHWDUL (IBM/360) :$1 0$1 /$1 /$1 0$1 /$1 /$1 /$1,QIRUPDWLFD 'LVWULEXLWD Il Personal Computer (Avvento del PC e dei Minicomputer) Nascita e boom delle reti (locali, geografiche, miste) (YROX]LRQHYHUVRVLVWHPLDSHUWL *HVWLRQH GHFHQWUDOL]]DWD Periferia eterogenea ed intelligente 1DVFLWDGHJOLVWDQGDUGper: Cablaggi strutturati Protocolli di Comunicazione

7 Sistem i Distribu iti: vantaggi e problem atiche &RQGLYLVLRQHGHOO,QIRUPD]LRQH &RQGLYLVLRQHGHOOH5LVRUVH $FFHVVRD5LVRUVH5HPRWH $OWD$IILGDELOLWj &UHVFLWD*UDGXDOH &RQYHQLHQ]D(FRQRPLFD PDLQIUDPH /DFRPXQLFD]LRQHWUDFRPSXWHUULFKLHGHVROX]LRQLWHFQLFKH FRPSOHVVHULJXDUGDQWLXQDVHULHGLSUREOHPL 5LFH]LRQHH7UDVPLVVLRQHILVLFD &RQWUROORGHJOLHUURUL &RQWUROORGLIOXVVR &RQYHUVLRQHGHLGDWL &ULWWRJUDILDHVLFXUH]]D 6LQFURQL]]D]LRQH 8QDSSURFFLRORJLFRqTXHOORGLDQDOL]]DUHWDOLSUREOHPDWLFKH VLQJRODUPHQWH ³'LYLGHHW,PSHUD

8 Modelli a strati (o a livelli) ¾ ¾ ¾ ¾ ¾ La suddivisione delle funzionalità secondo un modello a strati agevola la gestione della complessità Ciascuno VWUDWR (o OLYHOOR):» è responsabile di un sottoinsieme definito e limitato di compiti» funziona in maniera lascamente accoppiata con gli altri» interagisce solo con gli strati immediatamente superiore ed inferiore» fa affidamento sui servizi forniti dallo strato immediatamente inferiore» fornisce servizi allo strato immediatamente superiore Alcuni strati sono realizzati in software altri in hardware Vantaggi:» l indipendenza tra gli strati consente la sostituzione di uno strato con un altro di pari livello che offra i medesimi servizi allo strato superiore» limitare le funzionalità di uno strato ne semplifica la realizzazione Svantaggi:» L eccessivo numero di strati può portare ad inefficienze 9 6WUDWRQ 6WUDWRQ 6WUDWRQ 6WUDWRQ

9 Modelli di rete a strati a confronto User Netw. Appl. Session End to End Routing Data Link Physical Application Presentation Session Transport Network Data Link Physical Application Service Internetwork Network Transaction Service Presentation Service Data Flow Mana. Trans. Servi. Control Virtual Route Explicit Route Transm. Group Data Link Physical 26, '(&1(7 7&3,3 61$

10 Gli standard ¾Standard de iure» Ad es. ISO/OSI ¾Standard de facto» Ad es. TCP/IP

11 4XDQGRVWDQGDUGL]]DUH"

12 (Very) Brief History of the Internet ¾ 1961: Kleinrock - queueing theory shows effectiveness of packetswitching ¾ 1964: Baran - packet-switching in military nets ¾ DARPA (Defense Advanced Research Projects Agency) contracts with BBN (Bolt, Beranek & Newman) to create ARPAnet ¾ First five nodes: UCLA, Stanford, UC Santa Barbara, U of Utah, BBN ¾ 1974 TCP/IP specification by Vint Cerf (part of Berkeley UNIX) ¾ 1984 On January 1, the Internet with its 1000 hosts converts masses to using TCP/IP for its messaging (deployment of TCP/IP) ¾ Early 1990 s: ARPAnet decommissioned ¾ 1991: NSF lifts restrictions on commercial use of NSFnet (decommissioned, 1995)

13 Statistics from the IITF Rep ort, The Em erging Digital Econom y * ¾ To get a market of 50 Million People Participating: Radio took 38 years TV took 13 years Once it was open to the General Public, The Internet made to the 50 million person audience mark in just 4 years!!! ¾ Released on April 15, 1998 'HOLYHUHGWRWKH3UHVLGHQWDQGWKH863XEOLFRQ$SULOE\%LOO'DOH\ 6HFUHWDU\RI&RPPHUFHDQG&KDLUPDQRIWKH,QIRUPDWLRQ,QIUDVWUXFWXUH7DVN)RUFH

14 Internet ¾ Da infrastruttura militare ad infrastruttura di ricerca, ed infine ad infrastruttura commerciale (ma non solo ) ¾ Basata su TCP/IP» Standard de facto» Aperto ¾ Libera e Senza regole?

15 IETF, What is it? ¾ IETF, Internet Engineering Task Force ¾ An International Open, Non-Membership Organization ¾ It is Constituted under the ISOC (Internet Society) Legal Umbrella ¾ An IETF member is anybody who subscribe to one of the IETF mailing lists and take part in the discussion ¾ No subscription/acceptance procedures ¾ Many IETF members never went to an IETF meeting, nevertheless they make very useful contributions to the IETF work

16 IETF, What is its Scope? ¾ International Body to define,qwhuqhw6wdqgdugv, Informational and Procedural Documents ¾ IETF collaborates with the other International Standard Bodies (ITU, W3C, IEEE, ISO, ETSI...) to avoid overlaps and duplications ¾ IETF does NOT prepare Policy Regulations!» a domain name is what is defined in RFC 1035 ¾ IETF requires products interworking, but does not make nor endorse ANY product

17 IETF, The IETF Motto ³:HUHMHFWNLQJVSUHVLGHQWVDQG YRWLQJ:HEHOLHYHLQURXJK FRQVHQVXVDQGUXQQLQJFRGH

18 IETF, Working Grou ps ¾ Working groups, each with a specific charter and a specific scope and limited life:» produce technical specification on a specific topic» revise the technical specifications» control the actual implementations status and interworking» progress (or abandon) specifications to Standard» dissolve when work is finished

19 IETF, The RFCs ¾ RFC, Request for Comments ¾ NOT all RFCs are STANDARDS! ¾ Informational RFCs (for example BCPs) ¾ Experimental Status Specifications ¾ The Standard Track:» Proposed Standard» Draft Standard» Standard (Internet Standard no. xxx) ¾ Where to get them?» ftp://ftp.ietf.org/rfc/» ftp://ftp.nic.it/rfc/

20 ¾ Today around 30 people in the group ¾ Work spans 2 laboratories:» University of Napoli» CINI/ITEM a research lab of the Italian University Consortium in Computer Science & Engineering ¾ Funding mainly from EU, Industry, with some money from national and local government.

21 COMICS Lab: COMputers for Interaction and CommunicationS ¾ Research areas:» Traffic Measurement and Analysis» Network Monitoring» QoS in heterogeneous networks» Traffic Engineering» Wireless Mesh Networks» Management and control of network infrastructures SLA, SLS, Policy based management» Security, Reliability and Resiliency» Multimedia services engineering»

22 Progetti DIS/ Com ics ¾ Modello cooperativo tra docenti, dottorandi, studenti, collaboratori, aziende ¾ Creare competenze sulle tecnologie ¾ Creare valore aggiunto grazie alle competenze acquisite (non vendendo software!!! ) ¾ Piattaforme sviluppate sono su sourceforge oppure liberamente scaricabili dai nostri siti.

23 D-ITG, Distribu ted Internet Traffic Generator Traffic Generation and Measurement measure main QoS parameters (GHOD\MLWWHUSDFNHWORVVWKURXJKWSXW, ) in the presence of realistic traffic emulate network applications behavior test real implementations of novel protocols (,3Y6&73'&&3«) easily emulate very complex scenarios Distributed platform with multiple senders/receivers and remote logging Capable to emulate traffic statistics of real-world applications VoIP, DNS, Telnet, Network Games, Running on different architectures X86, ARM, XSCALE (Intel IXP) Windows, Linux, Montavista Linux,.. Currently used by several research projects all over the world VirginiaTech, Rutgers-Columbia-Princeton, Lucent Bell Labs, IBM Research, D-ITG,

24 D-ITG: some success stories ¾ A large number of Google Hits!!! ¾ Slax-tlc ¾ OSiAN Solution (an open source based network solution for African National Research a nd Education Networks. ¾ Orbit ¾ The Network Toolkit of CACETechnologies ¾ Traffic Generator of CISCO certifications learning systems ¾ D-ITG as a traffic generator for the NASA Crew Exploration Vehicle (CEV) Space communication Link sizing.

25 Hynetd: Hybrid Network Topology Discovery Topology Discovery: Fields of applicability: Fault isolation, performance analysis, network planning, node and service positioning, topology aware algorithms, network communications optimization, simulation studies ¾IP level» Routers, links, subnets ¾Hybrid approach (both active and passive) ¾High accuracy, short discovery time, low intrusiveness ¾Ongoing Work:» Porting onto embedded platforms (e.g., Intel IXP)» Distributed Discovery Planetlab ¾Hynetd,

26 Plab Measurement and Analysis of Network Traffic Extracting traffic properties at several levels By hosts, conversations, connections, flows,.. At SDFNHWOHYHO: inter-packet times and payload size Header inspection Payload inspection Traffic characterization and modeling, user/application profiling, anomaly detection,.. A flexible architecture: running on Linux, FreeBSD, MacOS-X capable to rapidly analyze traces of hundreds of millions packets enriched with several filtering capabilities (WFSGXPS-style, by payload, by time, ) ¾ Plab,

27 TIE: Traffic Id entification Engine ¾ A platform working as a multiple classifier system ¾ Purpose: to allow the community to work with VKDUHG WRROVDQGGDWD to investigate several aspects of traffic classification» Offline, Online, historical web reports» Easy to add: classification techniques, classification features, combination strategies» Programming API and documentation. Users mailing list.» Anonymized Traces with GT data» Code to the data

28 TIE: Traffic Id entification Engine ¾ TIE is a FRPPXQLW\RULHQWHGproject for traffic classification» Project started in Collaborations too! KWWSWLHFRPLFVXQLQDLW

29 IMS based services: CON FIAN CE & DCON ¾ ¾ CONFIANCE: CONFerencing IMS-enabled Architecture for Next-generation Communication Experience» Open source implementation of the IETF XCON Framework, compliant with the IMS specification Extends RFC 4353 Protocol-agnostic (not only SIP) Data Sharing (not only audio/video) Suite of Protocols: Conference Control (CCMP), Floor Control (BFCP), Call Signaling (SIP/H.323/IAX/etc.), Notification (Event Package)» DCON: Distributed Conferencing» An extension of CONFIANCE aimed at providing scalability through distribution of components/functionality» Based on XMPP-based spreading of information» Draft proposal from UoN at the IETF (waiting for time to become ripe )»

30 Med ia Control (IETF m ed iactrl WG) ¾ A framework for controlling a Media Server from an Application server» Similar to MRFC (Media Resource Function Controller) & MRFP (Media Resource Function Processor) approach in the IMS» Exploiting SIP (with so-called COMEDIA negotiation) to establish an ad-hoc mediactrl Control Channel» Making use of so-called control packages designed for specific applications (conferencing, IVR, etc.)» Currently under definition in the IETF: UoN among the most active contributors: z Provided an advanced open source prototype of the overall framework z See demo slides at the latest IETF meeting in Vancouver:» z Sourceforge web site:»

31 Conclu sioni ¾ La ricerca è aperta per definizione. ¾ Noi proponiamo non solo la condivisione di tool (aperti) ma anche quella dei dati. ¾ I paradossi.

32 Looking at the future: Hackers, Unite!!!

33 Reference e Fonti (u fficiali) ¾ Internet History and Growth, William F. Slater, III Chicago Chapter of the Internet Society September 2002 ¾ IETF In Italy History and Future View, Claudio Allocchio GARR

34 Domande?