ICTD2007: Session 4 - Alternatives to Real-time Internet Interactive Community Radio Revi Sterling, University of Colorado, Boulder Community Radio Observation that some ICTD projects improve access but exacerbate gender divisions. Social Science - women's development leads to community development Gender and Development in Radio * women trust radio * radio as "Africa's Internet" Radio - needs research but isn't a "sexy" technology Women * working, can't leave their environment * can't participate in radio b/c can't go to the radio station * but they want to be involved in radio * and radio wants to hear from them AIR Space * custom communications device to assist in women's participation in radio * "talk back" on radio airwaves - giving interactivity * research that tests several social and dev theories * ad hoc mesh network forms - and transmit from radio to radio to the station What AIR is not * it is not a cell phone - nor is it meant to be * take issue with "Cell phone as computers" for Africa - many places don't have cell phone coverage, and there is not incentive to implement coverage for low population regions * don't make it look like a phone - if it looks like one, then husbands will take it and sell it in Nairobi * not an exercise in gender mainstreaming - women in these communities are so isolated that they aren't ready * not a top down approach - participatory action research PAR "Action Cycle" Women and Comunity Radio * they listen to the radio all the time -> it's in their language * entertainment, escape, and school * some women place batteries higher than food! - will go hungry to get batteries * development radio programs -> formed women's listening clubs - but it takes them away from their duties - requires change in lifestyles - * took to the devices - wanting to create news - excited to be on the air and programming - initially thought they would make requests - but found that they are content producers as much as content requestors Push vs Pull * "Globalisation from below" - Massey 1994 * Sibanda 2005 quote AIR Hardware features * very low power ARM processor 7 day operation * simple push to talk operation * commodity 802.11 and flash ram usb devices for networking and storage * speech is amplified and passed through low freq bandpass filter * LEDs provied devic transmission status info * "Voicemail" Device Block Diagram - diagram Delay Tolerant Network Design * store and forward model: attempt to use the minimum number of redundant transmission to ohter devices to accomplish successful delivery * if all messages get through, parameters are relaxed * if messages dropped transmission parameter are made more aggressive * first send to three devices, until get to station * based on how long it takes for the data to get to the station, tune the number of neighboring devices sending to * acknowledgement of tranmission success via leds. Voice Routing * using fidelity comtech phocus array steerable phase array 802.11 antenna - sweep a semi-narrow beam across the area occupeid by community * person that goes to the market brings their device to carry the data to station range AIR Status * feasibility studies, baseline survesy, Search conference * Design validation * DVR pre-prototype * deployment in spring 2009 * India pilot * Interest in health and ag extension Offline Internet Access at Modem-speed Dialup Connections - Pakistan Umar Saif bandwidth available in Pakistan is 10-15 kbit range work is geared to deal with this bandwidth Motivation: Digital Divide * $40/month vs $4000/monh for 2Mbps * Average end user bandiwth is <10 kbps (100kpbs in US) Internet Pakistan * expensive international bandwidth * no real peering points * internet used over dial-up, prepaid "scratch card" provisioning * hard for ISP to predict Internet usage, b/c sold in prepaid card * almost never used for exchanging disseminatin accessing content Not a last mile problem * bottleneck is at the ISP * dsl link is 56 kbps * so soln is to do peer to peer exchange E-mail attachments * time oto exchange a 3.5 mb file on the internet - 1hr (16 kbps) Disruptive Tecnology * Of course Internet also started as an overlay over phone lines * a new kinds of internet * reminicent of pre-internet days Why is this practical? * Phone bills are becoming "Flat" - regional calling plans 199 rs/mont for free nationwide calls * But cannot always connect to the servier (ie. if in USA) * As long as one can identify a "close-by" host, "broadband" access is free * p2p systems already follow a similar model - e.g. incentive-based bittorrent Dialup p2p isp interleaving * use internet as a directory service, not a digital pipe * dial-up underlay * call host directly once you determine end-location Dialup bittorrent * BitTorrent in a sequential mode - bitTorrent normally uses 5 nodes at a time * Dialup doesn't work - you can only talk to one host at a time - find nodes which you can have the longest phone call - minimize calls (30 s setup latency for each call) Other challenges * How to figure out where the node are * overhead of peer connections: ~ 30 seconds * offline block discovery - last block problem * flask crowds - backoff for congestion contrl - everyone calling each other at once -> busy tones Peer connection overhead * worst case, need a call for each block * best case, you can get from one host Offline block discovery * virally spread knowledge of file blocks Last block problem * grab rare blocks first - based on directory determine which are "rare" blocks and connect first - this will also incentivize others to talk to you as well * favor those who will finish at the end of the connection - older guys forced to talk to younger guys * never choke due to bandwidth issues Flash crowds * each busy-ton costs 10 s * wait between calls (nominal) * Backoff in times of congestion (to many busy tones) * Optimum chocking interval graph - minumum wait period is 1s - network will prolly stabilize by 8 seconds Budget based download * get the rarest blocks and then wait for everyone to call you to get the remaining blocks * set up a budget for calls - and after that wait for people to call you Three evolving applications * p2p file sharring * web browsing * large email attachments dittorent.sourceforge.net umar@lums.edu.pk www.dritte.org (forum for researchers to discuss ideas for developing world) Warana Unwired: Mobile Phones replacing PCs in rural sugarcane (Rajesh's paper) Kentaro Toyama Agriculture in India * farmers with 1 acre of land, etc. Warana Nagar Warana Mharashta is a subdistrict... Warana Wired Village Project * state govt project * implemented August 1998 * used ftp, landline phone, to pc-enabled kiosks Original goals of WWVP * bring warana naga on NICNET (and internet) * create dataase of villagers on socio-economic growth * provide tele-education to both primary and higher educational Phase One: Initial Ethnography * is the porject actually achieveing what they want to do * rajesh spent 2 monts in Warana - observation, structured interviews, surveys - log kiosk usage statistics - attended meeting for kiosk operators * project started in 1998, rajesh there in 2005 - still active, used regularly by farmers Were goals being met? * most of the interesting things they want to do, they weren't doing Why not? * Technology - no local lanugage interface - connectivity is poor - dial-up at <10kbps, typically 4-6 - round trip time for data still two days * Infrastructural - Centers not designed for students, patients, etc * Social/political/economic - lack of budget to fulfill expectations - farmers need for internet/multimedia functionaly limited - interested in the kiosk, but didn't understand waht it was for - lack of training and qualified trainers Actual Use * seeing lakc of use, convert it to something that could be used * Internal Account MIS - register land - issue harvesting permits - self fertilizer through credit Mounting Challenges * High maintenance cost - at any time, a few in disrepair - dusty - open because they were alwasy fix it - stored pcs prone to rats chewing away wires * Intermittent power * Network flaky - low quality dialup The engineering Problem * Can we preserve the funcaltiy of the existing PC system with something cheaper and more reliable Turned PC-based Kiosks to Warana Unwired - sms-based mobile phones New Mobile-based Set-up Pilot Experiment * network coverage? * usability? - none of farmers used sms? * transition to a new system * ongoling operational cost? * Unanticipated issues? Implementation * October 2006 - smsm server toolkit * Implemente parallel mobile phone based system * Under auspices of cooperative's manageing director * Involved IT mgr at WWWP Data Collected * software logs of sms server and central database * survey of kioks (7 pilot, 7 non-pilot) * Interviews with kiosk operatoers and farmers Impleemntatio * server reboot once a month * running continuously since octoer 2006 * 24-hour access * 1238 unique farmer requests served in first three months * response time of sms on order of seconds Costs * New PC System 394 Rs.yer * SMS mobile 159 * sms mobile w.out kiosks 111 * GPRS (no skiosks, sms discount - 91 * provider noticed spike in usage and offered discount Other issues * Good: - advantage of mobiles - charge phones in tea shops, battery lasts several days * Bad - sms not suitable forall queries or data - land registrationnot implemented via SMS - too much input - data availablity dependent on server - minor error rates (3.2%) Photos of the system in action! Freed farmers to get information wherever they needed it! Farmer Responses * farmer for satve village (one of the pilots * inidal disbelief turns to excitement * test system with older information - then excited when older info was right! Farmer from neighboring village - we want it to Related work * Ict for agri: eSagu, e-Choupal aAQUA * advantage of modible phones - duncombe and heeks - donner 2997 - jensen 2007 * SMS based solutions Conclusion * first time replaced a PC based system with a mobiel based system * impractiacly expensive many-PC system replaced with an affordable single-PC system * initial investment of 50 million - not successful or cost effective - so can't replicate - if mobiel phones work perhaps can replicate! - rather than 50 * Is this development? - minor impact on farmers (savings of $10/yr) - cost savings to sugarcane cooperative - significant amount, but not huge impact on lives * Shouldn't overstate case for mobile phones - evidence that mobile phoens are fantastic - but can't generalize and say that mobile phoens are good for everything everywhere! * Future work: develop simple IT systems for agri cooperatives Q&A Q: (ken keniston) I did the first study in 2001, returned in 2005 Given all the materials about how much cool stuff there was there, and then realized what was going on. I gave up and Rajesh chose to stay Fertile area, sugarcane, dairy. One of few places with working cooperative. Very little problems between farmers and sugarcane factory. Govt' subcontracted a lot of project and abandoned when ran out of money, handing it to the farmers, Q: Did you consider the technical sustainability - maintenance sturcuter that you were planning to have for these kidn so fsystems A: (Revi) Device has a lifespan of 3 yrs. B/c using commodity hardware and there is a local community - hope that they can fix, etc. Hope that designed ruggedly and sustainably enough to last several years, Concerned that Q: Existing Handsets, or given? Did you look at corrollary usages on how they use the phones A: (Kentaro) Give kiosk operater a set of phones, but kiosk operator also taught farmers how to use the system on their own. So virally spread, and lots of calls from unidentified numbers Q: How much did it cost of interactive video machine? How do you provide that they wont take the radio machine away. How do you prevent the man to take the radio A: Device costs $100. At volume $5/device. Man taking device is the case of the phones. Men thought that there was no market value, were nonplussed. Interested in seeing development Q: (John Canny) If you have available other links between people using cable, or wireless point to point linkes. Is there incentive to create broadband internet infrastructure A: That's what we're are moving towards. Pakistant will see the world's largest WiMAX deployment. So there will be a pool of highbandwidth links. Q: (Sonesh Surana) to kentaro: For mobile phones being used, did you see unauthroized calls in the kiosk. A: I don't know/ Used an only-SMS package. Q: For revi (Paul Javid); In a lot of ways we like broadcast And you demo ways for people to talk back. Interested in use cases. Perhaps if ou introduce new experts (healthcare, etc) what do you think would happen. Interested in what you have done and what interested in doing A: (Revi) There are experts and governing boards involved, and had to come up wtih a "sharing plan". Situation when the divce could be given to hte police or other. Have not had experts araise. If they use the device to rant or talk to a friend, then it won't air, but if you give good content then it would be aired. So the station has improved quality b.c they can review the content Q: to kentaro but i missed it Q: For expenses for Warana project - for the entire piece in what you send through SMS - Q: (kentaro) On SMS - continuing discourse, replacement of mobile handsets with kiosk model... most of the informational domain is not just pricing etc, there is a lot of graphical interface with respecto farmers. Wh are we beomaining the fac thtat the farmers are not using the computers. If you look at institutional model - few farmers are marginalize. Kiosks serve many things besides internet access - short term credit, etc. When we look at models carefully, there are very few public places of this type available. There has been lobbing claiming unfair treatment of the govt on the quality of mobile servics A: It would be a mistake to say that all PCS should be replaced with a phone - there are many things that would be done. Mobile phones are not THE soln, platform must be matched to application. Warana - they were not that interested . Gov't expected an outcome that did not happen. The PC system in place was overkill for what they had there.