Indonesia .id Onno W. Purbo CONTENT In evaluating the Indonesian content produced over the Web, google search is used as a major tool. Indonesian content in different topics in Indonesian language as well as in English will be evaluated. More than ten (10) commonly found word in certain context are used as keywords to analyze the number of certain topics and certain language. A large number of URLs are found after Google search. The following tables will show the number of URLs found as well as rank each topics. The evaluation is based on the data obtained in year 2002 as well as 2003, and, thus, increase in the number of content can also be evaluated. It is interesting to notice a significantly different emphasis of content in Indonesian as compared to English. Table. Indonesian Content in Indonesian Language Rank Topics in Indonesian Language # URL 2002 # URL 2003 % Increase 1 Technology 1,788,210 2,745,100 53 2 News and current affairs 1,521,300 2,556,600 68 3 Education 1,176,300 1,603,330 36 4 Culture and literature 1,143,760 1,583,830 38 5 Industry/Business 1,123,800 1,978,300 76 6 Government 1,113,770 1,593,150 43 7 Civil society 1,059,290 1,768,300 66 8 Health/Nutrition 810,250 1,283,670 58 9 Commerce and tourism 725,047 1,279,612 76 10 Rural development 517,960 869,200 67 11 Political groupings 422,987 767,790 81 12 Agriculture 242,321 404,890 67 13 Non-government organizations 231,450 310,770 34 Table. Indonesian Content in English Rank Topics in English # URL 2002 # URL 2003 % Increase 1 Commerce and tourism 9,525,870 29,057,880 205 2 Industry/Business 8,571,900 27,520,000 221 3 Civil society 7,698,000 23,538,000 205 4 Government 5,979,990 18,509,200 209 5 Culture and literature 5,611,100 18,663,000 233 6 News and current affairs 5,395,100 17,472,900 224 7 Education 5,039,470 14,528,510 188 8 Technology 4,800,400 13,713,800 186 9 Political groupings 3,339,239 7,998,500 139 10 Health/Nutrition 2,608,250 8,766,800 236 11 Rural development 2,501,030 6,275,900 151 12 Non-government organizations 2,489,215 7,456,860 199 13 Agriculture 2,253,900 5,878,500 161 . As clearly shown in these two (2) tables, more emphasis on technology, news and current affairs, education, culture and literature are apparent in Indonesian language content. Consistent behaviour observed in the year 2003. While Indonesian content in English, which aim to reach a broader English audience on the Internet, put more emphasis on commerce and tourism, industry/business, civil society, and government. It is consistent in the year 2003. Thus, the interest, the needs of Indonesian readers seem to be different than that of English readers. In both Indonesian and English content, not many content produced in rural development, non-government organizations and agriculture. A significant difference apparent in the number of URLs between year 2002 and 2003. A significant increase in the range of 140%-240% in is apparent in Indonesian content in English as compared to Indonesian content in Indonesian language, that reaches only 40-76%. Thus, although there is an increase in both languages, increase in English content is much higher than the local Indonesian content. Ratio of Content in Indonesian Language relative to English is listed in the following table. % Indo 2002 % Indo 2003 Commerce and tourism 7.1% 4.2% Industry/Business 11.6% 6.7% Civil society 12.1% 7.0% Government 15.7% 7.9% Culture and literature 16.9% 7.8% News and current affairs 22.0% 12.8% Education 18.9% 9.9% Technology 27.1% 16.7% Political groupings 11.2% 8.8% Health/Nutrition 23.7% 12.8% Rural development 17.2% 12.2% Non-government organizations 8.5% 4.0% Agriculture 9.7% 6.4% In the year 2002, the average Indonesian content in Indonesian language contributes only about 15.3% of all content on Indonesia. In 2002, the percentage of Indonesian content in Indonesian language reaches its 27.1% peak in technology related areas; followed by health/nutrition at 23.7% and news current affairs at 22%. The Indonesian techies seem to get the most benefit from the network. In 2003, the ratio of Indonesian content in Indonesian language is much smaller than the Indonesian content in English. This is mainly due to significant increase in Indonesian content in English. The pattern remains the same technology content still dominating the Indonesian content. ONLINE SERVICES: E-COMMERCE / E-BUSINESS In Indonesia, having 13+ million cellular subscribers is a big advantage to run a service based on cellular features such as Short Message Services (SMS). Business based on SMS is flourishing in Indonesia. A cost of 10 cents US$ / SMS is quite affordable for most subscriber. Having 13+ million subscribers, it is quite common to see a rate of 20.000 SMS/day requesting for service. Thus, it is not surprising to see the provider to get around US$30-60.000/month for their services. PlaySMS http://playsms.sourceforge.net is an open source implementation of the key software to provide such service. A PC connected via data cable to a Nokia 6110 may run the service. A decent, linux box with Apache, MySQL and some PHP scripting would be sufficient to provide the service. DISTANCE EDUCATION & E-LEARNING The award winning http://www.ilmukomputer.com as "The Continental Best Practice Examples in the Category e-Learning" received in the last World Summit on Information Society (WSIS), Geneve, 9-12 December 2004, is one of the best Indonesian distance education & e-learning site on computer knowledge. Interestingly, management strategy on how to run a free e-learning site is described in http://ilmukomputer.com/populer/romi-elearning.php IlmuKomputer.COM is mostly (80%) accessed by Indonesian users. Some of webalizer data can be found at http://ilmukomputer.com/sponsorship.php http://www.ilmukomputer.com is lead by Romi Satria Wahono (romi@romisatriawahono.net). All about ilmukomputer.com, such as, introduction, summary of activities, accomplishments, working groups etc is described at http://ilmukomputer.com/project/romi-ilmukomputercom.php. More than 500 ebook, tutorials and articles are hosted at IlmuKomputer.Com. These knowledge are in various category, such as, introduction, general, polular, thesis, serial articles etc. They have not indexed all the manuscripts, but the estimate number of page is about 100,000 pages. There are about 130 contributors as listed in http://ilmukomputer.com/penulis.php. The system is run by a team of 26 person in various areas as described in http://ilmukomputer.com/pengurus.php. IlmuKomputer.COM is not only distributing their knowledge through the web, but also through CD media. There are about 200 people in various area in Indonesia acting as CD distributor. A complete list of these distribution channels is http://ilmukomputer.com/distributor/index.php As an education portal, two ways interaction plays an important role. A total of about 3000 subscribers are listed in IlmuKomputer.COM mailing list listed in http://ilmukomputer.com/milis.php. In addition, IlmuKomputer.COM is collaborating with about 15 elearning communities listed in http://ilmukomputer.com/jaringan.php. Web traffic is one of the major headace to managed. At the moment, there are 15 registered mirror serving IlmuKomputer.COM as listed in http://ilmukomputer.com/mirror.php. However, there are a significant numbers of unregistered mirror in various high schools and universities. E-CONFERENCE / FORUM / E-COMMUNITY It was all started in early September 2003 in Ottawa, Canada at Richard Fuchs's office at International Development Research Center (IDRC) eleventh floor. Richard Fuchs, Nancy Smyth and Onno W. Purbo were getting together and talked about the possibilities to build a more solid open source development communities in Indonesia. Next few months, after going though many stages & phase of community based development; and funded entirely by the communities. Indonesia today has its own community based discussion platform as well as several free servers or open source development. All of these activities are self-funded by the Indonesian communities. Groups.or.id the free mailing list discussion platform is running. As of 1 january 2004, total number of mailing list is 1541 with total unique e-mail subscriber of 34,410. To be honest, the number is too much lower then the claimed total Indonesian internet users. No one own the project / initiative, it is an initiative of the Indonesian people. The historical achievement can be traced back at the log book of Indonesian mailing server project at http://www.ictwatch.com/milisproject/. The first mailing list server is a Pentium III 600MHz. The machine is collective donated from many people, such as, Onno W. Purbo, Irwin Day, Bona Simandjuntak, Donny B.U., David Sudjiman (Leader of KPLI Jakarta), Rio Martin (ITENAS Student), M. Ichsan (Admin Visionnet), Nies Purwati (Secretary General Mastel), Rusmanto (Editor Infolinux), I Made Wiryana (Prof at Gunadarma), Mas Wigrantoro RS (Coordinator GIPI Indonesia), Heru Nugroho (Secretary General APJII), Anton (Manager of Telematics Indonesia), KPLI Jakarta, Nona and Anonymous. On 31 Oktober 2003. the mailing list server GOUPS.OR.ID was born! Work is currently underway to build various other servers, such as, free webmail / popmail server, development server, as well as many other server for open source communities. Currently there are three (3) servers in addition to the first Pentium III mailing list server. Most of the open source developers can now be find at developer@groups.or.id. The impact assessment of less than US$200 contribution with lots of man hours volunteering work is performed using pflogsumm that reveal many interesting statistical fact of groups.or.id after its 2+ months of operation. Grand Totals ------------ messages 50137 received 116186 delivered 25 forwarded 1603 deferred (14874 deferrals) 7032 bounced 1295 rejected (1%) 0 reject warnings 0 held 0 discarded (0%) 142197k bytes received 581m bytes delivered 1409 senders 319 sending hosts/domains 31369 recipients 3464 recipient hosts/domains As shown in the table, a significant amount of traffic has been generated and passed to the subscriber. A total traffic closed to 581Mbyte has been delivered. Host/Domain Summary: Message Delivery (top 20) sent cnt bytes defers avg dly max dly host/domain -------- ------- ------- ------- ------- ----------- 20630 101440k 550 1.0 m 48.9 m yahoo.com 14346 70691k 0 1.9 s 10.3 m groups.or.id 6152 17845k 0 10.4 s 4.3 m cbn.net.id 3878 20795k 262 1.9 m 1.1 h plasa.com 3656 16087k 892 9.1 m 22.1 h telkom.net 3472 16693k 258 2.1 m 1.4 h hotmail.com 3358 11252k 36 32.0 s 49.3 m indosat.net.id 2639 11705k 28 49.6 s 2.2 h indo.net.id 2007 8723k 0 8.7 s 17.4 m jsx.co.id 937 3874k 2 37.6 s 21.6 m ictwatch.com 895 3577k 3 15.6 s 29.4 m centrin.net.id 794 3625k 25 2.1 m 3.4 h dnet.net.id 787 3118k 10 41.2 s 41.3 m rad.net.id 681 3120k 928 1.0 h 9.3 h eudoramail.com 662 3194k 405 23.3 m 11.1 h lycos.com 636 2967k 10 1.3 m 32.9 m rantaunet.com 634 4437k 1 1.5 m 30.4 m ptcpi.com 633 4425k 0 53.0 s 14.5 m indosat.com 624 3347k 8 1.5 m 42.1 m yahoo.com.sg 591 4005k 45 2.3 m 42.0 m gmx.net Shown in the above table that yahoo.com is still number one (1) e-mail recepient followed by CBN, plasa.com, telkom.net, and hotmail.com. Note that yahoo.com, plasa.com, telkom.net, and hotmail.com are all free mail services. It seems many users who like to read / listen to discussion will likely to use free mail service. Host/Domain Summary: Messages Received (top 20) msg cnt bytes host/domain -------- ------- ----------- 42234 95492k groups.or.id 6431 39723k from=<> 190 924k returns.groups.yahoo.com 117 429260 yahoo.com 101 421521 cbn.net.id 86 360780 telkom.net 38 221984 indo.net.id 38 111771 indosat.net.id 33 350089 hotmail.com 24 86155 netscape.net 20 18055 imt.co.id 19 97322 informatika.org 19 74513 spinfinder.com 17 71022 ictwatch.com 15 68712 fmi.com 15 44844 arekkaj.org 14 152548 centrin.net.id 13 24644 dml.or.id 12 144792 jsx.co.id 11 99336 pks-jaksel.or.id A bit different from the recipent, the sender of mail to the mailing list server is still dominated by yahoo.com. However, the following hosts are CBN, telkom.net, IndoNet and IndosatNet. It seems quite a bit of the mailing list poster will like to pay for a more reliable e-mail service from the provider. Senders by message count ------------------------ 504 rantau-net-bounces@groups.or.id 415 proletar-bounces@groups.or.id 314 genetika-bounces@groups.or.id 182 nyambung-bounces@groups.or.id 168 kebumen-bounces@groups.or.id 138 huttaqi-bounces@groups.or.id 104 visibaru-bounces@groups.or.id 100 allp3i-bounces@groups.or.id 94 adacom-bounces@groups.or.id 87 labsaj-bounces@groups.or.id 85 fisika93-itb-bounces@groups.or.id 79 developer-bounces@groups.or.id 66 unhas-ml-bounces@groups.or.id 63 atekbl-bounces@groups.or.id 63 eramuslim-bounces@groups.or.id Shown in the table is the most active mailing list at groups.or.id. genetika@groups.or.id is known to be the mailing list where most Indonesian ICT activists meet & have a discussion. One of the most interesting phenomenon is kebumen@groups.or.id, Kebumen is a small city in mid Java island. It is amazing to see communities in such remote area get together on the Internet. The migrants originated from Kebumen in various cities initiate such discussion group. ICT INDUSTRIES AND SERVICES: Quite significant & consistent ICT industrial data can be found in the Indonesian Chamber of Commerce site (www.kadin.net.id/busisinessnet/) and the Indonesian Yellow Pages (www.yellowpages.co.id). In 2003, a total of 1608 Indonesian ICT companies can be easily searched through the Indonesian Yellowpages www.yellowpages.co.id. It is a significant (148%) increase from 2002. The summary is as follows # companies 2002 # companies 2003 % Increase computer consultants 53 60 13% computer internet 59 89 51% computer programming consultants 52 137 163% computer software 68 314 362% computer total solution 27 72 167% Multimedia 13 17 31% Software 88 148 68% e-commerce 2 0 -100% information technology 24 161 571% internet - services 133 358 169% internet data 13 22 69% internet portal 25 37 48% internet provider 90 174 93% web design 2 19 850% It is shown clearly that demand in Internet related business is quite high as reflected in the large increase in companies offering Internet services. Computer consultants, programming, software, as well as providing total solutions are also maturing to meet the demand. It is interesting to note that companies listed to provide e-commerce services are diminishing. It is a logical consequence of an untrustworthy business conduct by many Indonesian lamer. Stolen credit card used to buy goods over the Internet. It is a shame, the Indonesian police force as well as many communities is working on it. Based on the Indonesian Yellowpages data, the Indonesian ICT companies are distributed unevenly. Most of them are located in Jakarta, followed by Surabaya and Bandung, which is only 4 hours drive from Jakarta. Only few companies spread outside Jawa Island, namely, Medan (43), Bali (32), Palembang (4) and Makassar (6). City # companies 2002 # companies 2003 % increase Jakarta 466 1114 139% Surabaya 25 143 472% Medan 29 43 48% Bandung 75 120 60% Denpasar 15 32 113% Yogyakarta 14 76 443% Semarang 23 70 204% Palembang 2 4 100% Padang 0 0% Makassar 6 ? Total 649 1608 148% A significant increase (443%) is apparent in Jogyakarta. It is interesting to note that The Jogyakarta Sultan has declared to develop their area as cyber area. Note that Jogyakarta is also known as city of students with many universities and colleges, such quality of human resources is needed to back up a significant increase in IT business activities. INDONESIA INTERNET INFRASTRUCTURE The easiest way to probe the Indonesian Internet infrastructure development is through monitoring the expansion of Indonesian Internet Service Provide (ISP). Commercial ISP was started by IndoInternet, which is known as IndoNet http://www.indo.net.id in 1994. Most of the commercial Indonesian Internet infrastructure can be investigated through the Indonesian Internet Service Provider Association also known as Asosiasi Penyelenggara Jasa Internet Indonesia (APJII) http://www.apjii.or.id. As of 3rd quarter 2003, there are * 124 Member APJII * 93 ISP with operational licensed holder * 73 ISP connected to the IIX * in 90+ cities, all provinces as listed http://www.apjii.or.id/layanan/lokasi.php?lang=ind APJII provide common facilities, such as, * APJII Indonesia Internet Exchange (IIX) * APJII IDNIC * Domain Registration & NIR (APNIC) In reality, various form of Internet cafes or neighborhood network serve 60-70% Indonesian Internet users. It is interesting to note that all of these activities are privately driven and not publicly driven. Based on the annual report of the Indonesian ISP Association (APJII) that can be downloaded from http://www.apjii.or.id/dokumentasi/statistik.php?lang=eng, the estimated Internet users and subscribers up to the end of 2002 is as follows, Subscribers Users 1998 134.000 512.000 1999 256.000 1.000.000 2000 400.000 1.900.000 2001 581.000 4.200.000 2002 667.002 4.500.000 2003* 800.000 7.550.000 Table: Growth Indonesian Internet Subscribers and Users *Estimated up to the end of 2002 Source: APJII (www.apjii.or.id) It is clearly shown that a doubling of Internet users is common growth rate in Indonesia. In 3rd quarter 2003, APJII estimated that the total Indonesian Internet users are around four 7.5 million users. More detail study on its profile can be found in the research done by APJII. It is common to see a majority of male, young (25-35 years old) and educated people (high school graduate, university students or young professionals) would compose the majority of the Indonesian Internet users. Although, Indonesian would prefer to use international domain naming, such as, .com and .org, due to its flexibility and easiness in getting the name. A growth in new domain registered in the Indonesian top-level domain ID-TLD is apparent. New Domain Total Domain 1998 1.479 1.479 1999 2.148 3.627 2000 4.219 7.846 2001 3.478 11.324 2002 3.183 14.507 2003* 3.764 18.271 Table : Indonesian Domain Source: www.idnic.net.id In the year 2001, there is a decreased in new domain. Budi Rahardjo, the ID-TLD, argues that the reduction is mainly due to global dotcom failure worldwide. Budi is not commenting on the use of .com or .org on Indonesian sites. The allocated public IP address and AS number in Indonesia is growing as clearly shown in the table. Ipv4 Accumulative (blocks) Ipv6 Accumulative (blocks) AS Number 1999 256 3 2000 1.072 16 2001 1.553 29 2002 2.455 59 2003 2.377 131.073 69 Table : Increase in IP address and AS Number. *up to 3rd quarter 2003 Source: APJII (www.apjii.or.id) To reduce the international Internet traffic, APJII sets several Indonesia Internet Exchanges (IIX) in Jakarta, e.g., in Cyber Building in Mampang, Jakarta, in Kuningan, Jakarta and in Telkom Building, Jalan Gatot Subroto, Jakarta. The one at Telkom Building will likely to cease operation, as Telkom will terminate their lease. These IIX are interconnecting with each other. The IIX serves the connection of all ISPs in Jakarta at zero interconnection charges. The same principle is currently being pursued in various cities in Indonesia, such as, Surabaya, Bandung and Jogyakarta. It is needed, as most of intra-city traffic is now routed through Jakarta. From MRTG (Multi Router Traffic Graph) reports administered by Johan Alam (johar@the.net.id) the IIX administrator; we will find a significant number of increase in peak bandwidth through the IIX. Peak (Mbps) Feb '99 2.05 Jan '00 3.07 May '01 40.96 March '02 245.76 Sep '02 255.4 March '03 620.595 Jan '04 1.2Gbps Table: Peak IIX Traffic Source: IIX (www.iix.net.id) By analyzing the above traffic one may find that the traffic increased is mainly due to increase in corporate subscribers. In addition to it, Indonesian Internet users find more interesting Internet applications that require high bandwidth, such as, transferring MP3 files, multimedia, and online gaming over the network. As reported by Johar Alam (johar@the.net.id) the administrator of Indonesian Internet Exchange (admin@iix.net.id) and Heru Nugroho (hn@apjii.or.id), in early January 2004, the total IIX in country peak bandwidth is in access of 1.2Gbps. Since the international traffic is normally about three (3) times of local bandwidth, the peak Indonesian international bandwidth is estimated about 3Gbps. The peak bandwidth is normally about 70-80% of the maximum bandwidth. Thus, it is safe to estimate a maximum bandwidth of 2Gbps from Indonesia to the Internet The ratio of in-coming and out-going Internet traffic volume is about 1:10 as Indonesian is still consume more information rather than produce information. As reported by APJII, the table shows clearly that the government grants large number of ISP licenses. Thus, it seems no restriction is applied and only the professional would be able to show its capacity to operate an ISP and receive the operational license. ISP NAP Multimedia 1994 2 1995 18 1996 37 1997 43 1998 44 1999 50 8 2000 139 5 18 2001 172 16 24 2002 180 18 24 2003* 190 29 24 Table: Lisence provided by DIRJEN POSTEL *up to 3rd quarter of 2003. Source: APJII (www.apjii.or.id) Of the above licensed companies, some may apply for membership at the Indonesian ISP Association APJII www.apjii.or.id. As shown on the table, there are about 125 member of APJII as of the 3rd quarter of 2003. Not all licensed companies will joined the Indonesian ISP Association (APJII). APJII member is growing from 41 members up to around 124 in 2003. Shown in the table is the growth of APJII member by type of license. 1999 2000 2001 2002 2003* ISP 41 74 105 119 111 NAP - 1 3 6 6 Multi-Media - 3 5 7 4 Misc** - - 2 3 3 Total 41 78 115 135 124 Table: Growth of APJII Member *up to the 3rd quarter 2003 ** Wireless Internet & Internet for Research Education (IPTEKNET) Source: APJII (www.apjii.or.id) Not all APJII members require services, such as, IP address allocation and interconnection to the Indonesian Internet Exchange. Current status of APJII member is shown in the table. 1999 2000 2001 2002 2003* Member 41 78 115 135 124 Operational 35 63 82 90 93 Connected to IIX 12 24 49 70 73 Table: APJII Member Status * up to 3st quarter 2003 Source: APJII (www.apjii.or.id) EXAMPLES OF INNOVATIVE AND KEY INITIATIVES: INDONESIAN VOIP MAVERICKNET Virtually declared in cyberspace at genetika@yahoogroups.com, Sunday, 12 January 2003 10:19:39 GMT+0700, the Indonesian VoIP Maverick Network also known as VoIP Merdeka in Indonesian language officially run. VoIP MaverickNet is the result of reaction of Government plan to increase the phone tariff per first january 2003. VoIP MaverickNet is a community based VoIP Network run on top the Indonesian Internet Infrastructure. It is a free VoIP network & one of the most complex countrywide Internet Telephony infrastructures ever implemented. An open source gatekeeper (www.gnugk.org) is used as the main gatekeeper software. The VoIP MaverickNet archirecture is based on a VoIP gatekeeper cloud that acts as the main VoIP switch. In the Figure, it is shown the complicated gatekeeper structure of Indonesian Internet Telephony MaverickNet. RootGK - it normally runs as the highest-level gatekeeper in the infrastructure. In the Indonesian Internet Telephony MaverickNet, it is located at the Indonesian Internet Exchange (IIX). The second level gatekeeper is a stand-alone gatekeeper in the Internet to be used by public. It is basically a public gatekeeper, some calls it an Operator Gatekeeper (OGK). The third level gatekeeper is the gatekeeper located at the Proxy Server. It is basically a proxy gatekeeper, some calls it Local Gatekeeper (LGK). Please note how the numbering is set. An endpoint (EP) with a number of 62888881234567 must register to a public gatekeeper that handles prefix 628888. If the EP 62888881234567 registers to other public gatekeeper, other EP will not be able to call EP 62888881234567. At the moment this manuscript is written, there are a couple public major gatekeepers serving the Indonesian Free Internet Telephony MaverickNet, namely, Two (2) machines act as the RootGK located at 202.155.39.157 and 202.53.224.172. A couple more gatekeepers acting as the public gatekeeper serving a specific prefix, such as, 202.155.39.157 - for area code 62 88 (Root Gatekeeper). 218.100.4.194 - for area code 62 88 888 202.43.162.189 - for area code 62 88 999 202.150.8.15 - for area code 62 88 925 Some of friends in Indonesian Free Internet Telephony MaverickNet are donating their gatekeeper for public usage. Some of them can be reviewed at http://gk.vision.net.id and http://voipmerdeka.net/gkregistration/. A total of 147 Regional / Operator Gatekeeper can be found. There are unknown numbers of gatekeeper running at many proxy servers. Typical daily traffic is about 2000+ calls per day. Serving thousand of users. It is interesting to note that some users may have call duration in several days. Effort is currently underway to integrate the network into other international gatekeeper network. The integration is done through gatekeeper Neighbors settings. There are several active gatekeeper neighbors, such as, Singapore, Japan, England, Germany, Canada and Sudan. Note that these neighbors are not the formal Telco gatekeepers. The volunteers in these countries run it. Community education is the basic strategy in building such free & low cost VoIP network.. Mailing list is used as the main vehicle to do the education processes at voipmerdeka@yahoogroups.com and gk-admin@yahoogroups.com. ENABLING POLICIES At the time this manuscript is written, a fierce fighting is currently underway over several mailing lists on the Internet, such as, indowli@yahoogroups.com, genetika@groups.or.id, asosiasi-warnet@yahoogroups.com, telematika@yahoogroups.com, mastel-anggota@yahoogroups.com, between many individuals from the Indonesian grass root communities and individuals from the telecommunication regulators. The fight is on liberating WiFi from any requirement of government licenses. In October 2003, a draft on ministry of telecommunication decree on WiFi is being written. Many people from the communities fight for freedom, while some people on the telecommunication regulator insisted to license the band. Would it be fair to charge US$250 per year per Access Point for frequency usage, while the price of an Access Point is only US$100-150 each? Unfortunately, the government limits the license to operate the equipment only for ISP. Those who run the system without any license will be revoked. Such approach is claimed to bridge the digital divide and increasing the Internet penetration in Indonesia. As knowledge on WiFi deployment can be freely downloaded from the Internet at http://sandbox.bellanet.org/~onno/ and http://www.apjii.or.id/onno/. It is getting more difficult for the government to fight its own people as more people can build, install and run the system. As of end 2003, there is 5000+ outdoor WiFi installations, at 200-300 new additional installations per month, The government will have to face a large mass of people to fight with. It is a shame some the telecommunication regulator does not realize the impact of liberating 2.4GHz & 5 GHz WiFi band. An underestimate calculation reviles that there will be an increase from currently, 1. One (1) million to 17.8 million WiFi based Internet users. 2. Rp. 1 Billion/year to Rp. 21 Billion/year ISP tax. 3. Rp. 7 Billion/year to Rp. 128 Billion/year VAT. 4. Rp. 18 Billion to Rp. 600 Billion investment tax. 5. 50.000 units to 2 million unit computers. 6. 5.5000 units to 130.000 units ISM band equipment. 7. Such increase in ISM band equipment needs will justify an industrial migration from antenna and tower manufacturing to ISM band equipment manufacturing at US$4.5 million value. Not to mention spin off effect on many economical benefit to the country. There are many version of WiFi band liberation scheme; the one that proposed by Onno W. Purbo (and supported by many individual in the community) to the ministry of telecommunication is as follows, 1. Liberate ISM & UNII (2.4 GHz, 5.2GHz & 5.8GHz) band 2. No license and registration requirement for ISM & UNII band users. 3. No type approval requirement for equipment approved by FCC and ETSI. 4. Limit the maximum Effective Isotropic Radiated Power (EIRP) to 30-36 dBm. Those who violate the limit may be prosecuted under Telecommunication Act No. 36/1999. 5. The community coordinates frequency sharing and reuse coordination for Metropolitan Area Network (MAN). COMMUNITY BASED ICT INFRASTRUCTURE Let's play with the idea what if no Telco is necessary? What if no government & world bank involvement necessary in the development & building of an infrastructure in developing countries? What if it costs only 50 cents/students/month to get such infrastructure for schools in developing countries? Sounds like a dream for those who live in any development countries like I do in Indonesia. Fortunately, in reality, it can be easily done. It is not the equipment, nor the regulation, nor the investment that counts, it is the ability to educate critical mass that gain the required information & knowledge critical to the establishment of such infrastructure. Let's deal with education issue later and look at the community based infrastructure first. Shown in the figure is the typical self-finance bottom-up community based telecommunication infrastructure deployed in Indonesia. It has been deployed in several neighborhood as well as many offices to interconnect their branch offices over VoIP. It relies heavily on community based Internet infrastructure. The basic components of the community based telecommunication infrastructure are, Internet infrastructure Wireless Internet (WiFi) equipments, especially to bypass Telco's last mile. Internet Telephony Gateway PBX Telephone cables and handset runs throughout the neighborhood. Most the equipments are off the shelf gadgets, and fairly user-friendly and easy to use. Our experience shows that investment within a neighborhood network would be around US$80-100 to get a house 24-hour connection to the Internet. One still has to buy a computer to use the network though. Adding external antenna into WiFi equipments in 2.4GHz and 5.8GHz extend the coverage of the equipment to typically 5-8 km from the original 100-200 meter. Carefully design the Wireless Metropolitan Area Network (WMAN) by the community may then enable a citywide area network. Detailed techniques of such configuration can be freely downloaded from http://sandbox.bellanet.org/~onno/the-guide/wifi. Practical know how discussion is normally conducted at wifi4d@dgroups.org. Voice based communication is of particular interest in many rural communities as well as some urban communities where read and write culture is not widely spread. In addition having a 24 hours Internet connection in most modern offices would likely be a significant incentive to be able to interconnect their PBX in offices for long distance as well as local calls over their existing 24 hours Internet connection. A more complex community VoIP infrastructure may be suited the community needs. An example of a countrywide community based VoIP is the VoIP Merdeka (VoIP MaverickNet) http://voipmerdeka.net in Indonesia. It relies on hundreds of gatekeepers, based on the open source www.gnugk.org, to act as the main VoIP switches on the network.. Carefully design the numbering plan may enable a Telco-like infrastructure in calling each other. For example, the Indonesian community VoIP Network has its own call area of +6288. Interconnection to conventional Telco network can be easily performed using Internet Telephony gateway (ITG). In office complex, hotel, neighborhood, one may interconnect the existing PBX into VoIP infrastructure through the ITG. Detailed practical techniques can be freely downloaded from http://sandbox.bellanet.org/~onno/the-guide/voip/. COST OF AN APPROPRIATE ICT TECHNOLOGY FOR THE ASIA PACIFIC I would argue that cost-wise a community based infrastructure rely on WiFi, VoIP and PBX would have a much less investment and operating costs as compared to conventional cellular or public switch telephony network infrastructure. Thus, it would be of interest to look at such alternative infrastructure in developing countries. The cost reference here is the typical cost in Indonesia, a developing nation, in the Asia Pacific. In developed nations, the cost might be much lower. However, in some African countries, the cost might be higher. Not to mention some regulatory complexity. A typical Internet Telephony Gateway (ITG) with four (4) ports would cost about US$400 each. The ITG will translate conventional voice traffic into VoIP traffic over Internet infrastructure. To reduce the investment and operating cost, we normally connect the ITG to a Private Branch Exchange (PBX) and share the VoIP lines with more people in the office or neighbors. Thus, it is normal to connect all four (4) ITG's port to PBX. A typical PBX, such as, Panasonic KX-TA616 has 6 CO lines and 16 extensions. Panasonic KX-TA616 may cost around US$400-500. A ready to use client side WiFi equipment would cost about US$150-200. An RJ-45 UTP connector is ready to use on the back for this equipment to be connected to other network equipment on our local network. To extend the distance, a US$50-100 external antenna may be added. Such configuration may reach 5-8 km distances, a good enough distance to cover most medium size cities in developing countries. For those who like to build your own PC based client WiFi equipment may get a much-reduced WiFi investment. Total investment cost is about a maximum of US$1200 for WiFi equipment, Internet Telephony Gateway and PBX. The 16 subscribers connected through the extension of the PABX must pay these investment costs. Thus, investment cost per subscriber is about US$75. A typical cellular investment / subscriber is about US$300-400/subscriber. While, a typical fixed wire telephony network / public switch telephone network requires about US$1000/subscriber. Thus, WiFi & VoIP based infrastructure per line is much less as compared to typical investment per subscriber in cellular and in typical public switch telephone network. Based on the IP bandwidth calculation http://www.erlang.com, a typical full duplex G.723.1 (ACELP) 5.3Kbps compressed voice including the TCP/IP protocol header will require about 64Kbps for a four (4) lines VoIP. A dedicated 64Kbps Internet link would normally sell at US$350-400 /month directly from the Internet Service Provider (ISP). A no cost 24-hour local connection to the ISP can be done using WiFi equipment replacing the local Telco's leased line. Thus, the actual monthly operating cost per subscriber of such community based infrastructure is only US$350-400/16 or about US$20-25/month/subscriber including "unlimited" long distance calls over the Internet to other community based network. If US$20-25/month/subscriber seems too high, one can always resell their access to others in their communities as demonstrated by the Women in Bangladesh who subscribe to Grameen Phone. Both low investment cost and low operating cost has been demonstrated in a simple community based infrastructure based on WiFi, VoIP and PBX. Hope to see such infrastructure can be flourishing in many countries in years to come. EXPERIENCES IN INITIATING COMMUNITY BASED ICT MOVEMENT A clear result of a community based ICT development has been clearly shown in the development of Indonesian open source development community, e.g., groups.or.id, or the VoIP MaverickNet. No support from the government and no loan from World Bank & IMF are necessary. It is an art in itself to be able to create a countrywide community based self-finance self-motivated ICT movement. In this section, practical strategies in initiating such community based ICT movement will be described. There are several important ingredients that will enable such movement, i.e., 1. There must be a good cause (good reason for such moment). Some may see it as the vision; some may see it as the stumbling block, the challenge. 2. There must be a good informal leader that has the vision, knows the way, shows the way and goes the way. It would be the most difficult one to find such leader. 3. Ability to shift the paradigm of the common, from being a information / knowledge consumer into information / knowledge producer. Closely examine the technology trend; one may easily notice that technology is becoming more affordable and user-friendlier. Thus, a challenge in regulatory framework is apparent as shown in the table. Conventional Future? Technology Expensive & Complex Affordable & User-friendly Specialized Skill Yes No From Licensed by Government Common People Invest by Investor (Multinational) Common People Build by Vendor & Contractor Common People Run by Operator Common People For User & Customer Common People It is not surprising to see shift in the industry structure, such as, * Internet Café resells their bandwidth to their surrounding neighborhood and initiate a neighborhood network. * VoIP integrates PBX of the main office to its branch offices and bypass any long distance charges. * Offices, Neighborhood and communities bypass any Telco last mile infrastructure using low cost WiFi based equipments. * Cost of VSAT equipments is getting lower, in the range of US$1500-2000/ground stations. Not surprising to see some will run a VSAT infrastructure service based outside the country. Current advancement in ICT technology enables communities, offices, and resellers to act as unlicensed service providers and / or unlicensed physical network operators. Such shift creates a significant problem in the policy & regulatory framework that accept only licensed operators to run the network. Thus, no room is available for a community-based infrastructure; with homebrew equipments, from people, by people, for people. No room for WiFi based last mile. No room for VoIP based community network. In this example, the reason & challenge is very much clear. However, having access to low cost technology on hand is not enough. ICT knowledge in local language is fairly limited; translating the knowledge on the Internet or various technical books requires a significant amount of effort and funding. Thus, ability to generate and distribute ICT knowledge in local language at low cost would be the key. To sustain a digital divide bridge deployment in a long period of time, ability to self-finance local knowledge generation cycle is crucial. Key success in deploying the information infrastructure would rely heavily on the mass of quality & skill of the human resources. Ability to distribute the needed knowledge & skill at low cost would be crucial in creating demand & needed skilled technician to deploy the infrastructure. Ability to access ICT knowledge in English would be a plus. Interestingly, as equipment can be shared and is getting cheaper, funding is not the primary concern in deploying ICT infrastructure. In creating a self-finance knowledge cycle, there are basically two (2) critical strategies, namely, * Transforming some locals into local knowledge producer. * Supply created-by-demand strategy. Unfortunately, the whole process takes at least ten (10) years. Thus, never ever expect financial benefit in a short period of time. It will never happen that way. Most failure is likely due to demand created-by-supply path or heavy top-down approaches. Demand in information infrastructure as well as ability to exploit abundant information & knowledge on the information infrastructure may only occur in an educated society. Failure to increase the society's level of education will likely to impede the development of information infrastructure within the society. A conservative route would go for connecting the school and university network first, as it will affordably cost US$ 0.50/student/month with Return of Investment in 1-2 years. Furthermore, young generation would be an effective agent of change for their parents and surrounding communities. In developing country, like Indonesia, practical know how on various ICT technologies in local language is now in high demand. Publishers are begging for manuscripts on various ICT practical guides in local language. Most of the best Authors are in their 30's and acquiring their ICT knowledge through their school and university's Internet access. Thus, it is very crucial to integrate schools and universities to the Internet, as they are the one that produce the desperately needed local knowledge 3-5 years after their initial introduction to the Internet. Simplified Steps in Knowledge Generation & Demand Creation * Create discussion platform for 2 ways interaction. Such platform is basically to facilitate tacit (implicit) knowledge exchange. It may be radio or TV talk shows, or Internet mailing lists at no cost. Unlike most assumed condition by 1st world countries, no abundant local content is necessary to start with. Community may create their own content through community implicit knowledge exchange platform. Existence of informal community leaders is crucial in such platform, a leader not necessary the moderator. * It may take 1-2 years, before some individuals start analyze and synthesis the collective community knowledge. These individuals, mostly young, should make their knowledge implicit in written form. It should be distributed through conventional channels, such as, newspaper, magazine, radio & television, and, thus, transform the mindset of others who are not yet using ICT. It costs US$ 1-2 to access the knowledge in books or magazines. The young authors, in return, normally receive an incentive of US$ 15-25 /article or US$ 400-700 /book. The best selling and the most productive Authors, it would be easy to get equipment donation from various vendors. * Some may need physical contact through seminar and workshop to be convinced. We are normally looking at 500-1000 participants/seminar. The favorite seminar agenda would be to provide tutorial and demonstration on various ICT technologies. It costs US$ 3 /person to enter such seminars as many vendors may likely give sponsorship. Marketing through local newspaper or computer magazine is crucial in getting such large participants. A hands-on workshop with two (2) persons on one computer will be more costly about US$ 40-50 for 1-2 days workshop, and normally accommodate 30-50 participants. * As more people convinced, demand is created within the communities for their very own "digital divide bridge". Sound business plan, with US$ 2000-4000 investment and return in 1-2 years, will persuade people to invest their money to deploy their very own "digital divide bridge". A 50 cents/student/month in a school network or US$15-30/house/month with Return of Investment within 1-2 years is a good enough incentive for people to invest their own money into the "digital divide bridge". Success stories & word of mouth is the typical process to spread the words. Manufacturers and business communities will gladly responds to such opportunities. Some individual would spin-off and create their service companies to meet the demand. A conservative path would connect the school network first before connecting the neighborhood. * Deploy & maintain the "digital divide bridge" may be done at 50 cents/student/month at school or about US$ 15-30/month/house. Ah, we finally have financially sustainable digital divide bridges with minimal support from the Bank, the government & hopefully license free. In the end, as more and more people connected to the Internet. As shown clearly, all the sequences are self-finance. Having convinced, community will invest their own money into the infrastructure, run the infrastructure for their own community. These steps would not be completed over night. It will take years to complete these steps. Committed leaders & personals are needed for such long-term deployment ICT movement. To give some extreme ideas, we can actually mingle with the current education system. For example, * When I teach at Institute of Technology Bandung, I removed the necessary attendance. No mid term exam. No final exam. Mark is based on student ability to become a knowledge producer. A simple tactic, forcing the students to publish article in the magazine, e.g. 5 articles A, 4 articles B, 3 articles C etc. or publish a book for an A. * Other way to encourage the students, I give free access to the Internet for those who like help maintain and run the network system in my laboratory. In the end, I ask them to write book based on the knowledge gain during the process. The private sector will be gladly helping us in performing the process as we create huge demand for their products. Getting free equipments to do the experiments is not that difficult. Most companies would be interested in sponsoring the seminar / workshop if there are 500-1000 people showed up. Through these simple processes, we force young students to become a knowledge producer to benefit the whole society through local knowledge accumulation and publishing the knowledge. There is one important thing missing on the scheme, to find a committed leader that can lead the bottom-up community based ICT movement. It would be the most difficult task. However, one may find it quite easy to find one after having discussion / interaction with the communities for some years.