Ilaria Cianci, Luigi Alfredo Grieco, Gennaro Boggia DEE - Dipartimento di Elettrotecnica ed Elettronica, Politecnico di Bari

Via Orabona , 4 - 70125 Bari (Italy)

CCNx Meeting 12 -13 September 2012

.

Different Technologies Contents

Internet of today

Current architecture

based on NODES location-based routing (IP

addresses) network congestion weak support for mobility and

broadcasting

Internet of tomorrow

based on CONTENTS named-based routing improvement of network

performance good support for mobility

and broadcasting

Future Internet Architecture Content Centric Networking

Contents exchanging independently of physical location of hosts Distributed caching adoption Named-based routing (hierarchical structure of

names) Secure data sharing Reduction of servers computational load and

bandwidth consumption

CCN allows nodes to communicate based on

what data they need, without knowing a specific path to reach a specific node

CCN guarantees energy efficiency CCN routing avoids routing loops Nodes can use application data names directly

to forward Interests and data packets among each other

CCN suitability in Wireless Ad Hoc Networks

Collections of wireless nodes that can self-organize an arbitrary network without any infrastructure

Each node can communicate directly to

each other

Noisy, limited-range, and not secure wireless transmission Energy constraints

CCN Joker is a new fully customizable and open-source emulation platform that allows the emulations of basic aspects of a CCN node through a lightweight high-level program

It has been designed to run on top of wireless devices with limited resources in order to test the CCN exploitation in wireless ad hoc network environment

CCN-Joker

EFIKAs EFIKAs are small but complete and powerful computers based on a RISC PowerPC processor. These devices target mostly embedded applications and they have a very low power consumption

ATX board (153x118x38 mm); Freescale MPC5200B PowerPC SoC up to 466 MHz; 128 MB DDR RAM @ 266 MHz; 44 pin IDE connector; 1 PCI (33/66 MHz PCI 2.2); 10/100 Mbit/s Ethernet (Realtek 8201 Phyceiver); 2 USB ports (1.1); 1 RS232 Serial port D-SUB9; Stereo audio out, microphone and line input S/PDIF; Infrared port (IRDA) from 2400 bps to 4 Mbps; RTC clock (power management on/off).

The core module of CCN-Joker is an UDP server, namely JokerServer, listening on the port 9700. It is active on all nodes of the overlay and its role is to handle incoming requests generated by other CCN-Joker modules

UDP  PACKET

SOURCEPORT 9700

LENGTH CHECKSUM

DATA

JOKERSERVER

Which Source Port?

It is a JokerClientrequest. If the required contents are not in the CS or in the Repository, there will be generated

the corresponding Interest packets.

9705

9700

YES NO

REQUESTMANAGER

It is a Data Packet?

YESNO

CONTENTMANAGER

It is discarded.

9706

It is a CacheManager request. The list of all available contents in

the CS are sent in response to the CacheManager.

It is an Interest Packet?

Internal requests

External requests

Work-in-progress: Flow Control in CCN-Joker

Testbed

six EFIKAs equipped with 4GByte Flash

memory and with a Wifi module

Linux operating system

(OpenSuse) installed on top of

each device

When CCN-Joker daemon starts,

each Efika behaves as a

CCN node

E-mail: i.cianci@poliba.it Personal e-mail: ilaria.cianci@gmail.com

Telephone number: +39 080 5963301 Fax: +39 080 5963410

Skype contact: ilaria.cianci87 Web page: http://telematics.poliba.it/cianci/

CONTACTS:

I Cianci, L.A. Grieco, and G. Boggia," CCN - Java Opensource Kit EmulatoR for Wireless Ad Hoc Networks", Proc. of ACM Int. Conf. on Future Internet Technologies, CFI, Seoul, Korea, Sep., 2012.

Scenario

Preliminary Experiments Adopted parameters

Investigated metrics

Number of nodes = 6 Cache size = 125 [item] Catalog size = 2500 [item] Content size = 10 kB

request/s Experiments duration = 5 hours

Average number of content copies Hit ratio Average download time

copies in the network compared with a proportional trend

Number of content copies for the most popular content during the time

Average hit ratio for each content calculated as the ratio between the hit events and the total number of requests

Average download time: average time between when a request starts from a given node and when the required data is obtained

timet1 t2 t3 t4

W D1

I2 I3 I4I1 I5

timeout

D3

I2

D4 D2

I6 I7