Peering Wars

This was actually a quite old story issued by Renesys several months ago and it was also presented at NANOG 43 (June 2008). The story took a case of network-depeering and the actors were Cogent and Telia. At that time, both Cogent and Telia are considered as "near" Tier-1 providers. Why "near"?

• Cogent gets (got?) transit from NTT (formerly Verio) to reach AOL (ATDN?).
• Telia appeared to get transit from Verizon to reach certain networks.

Additionally, it is worth to mention that

• Tier-1 providers act like a cartel and have no incentive to add members
• "Near" Tier-1 providers can try to buy their way into the club (via "paid peering")

Renesys said that Cogent and Telia want to be "real" Tier-1, and from this point, the story began:

• On 27/Feb/2008 Renesys stopped seeing evidence of Telia's transit routes and Renesys promotes Telia to "Tier-1" (However, it could still be paying for some of these interconnections)
• On 13/03/2008 Cogent depeers Telia, claiming breach of contract; Renesys observes routes from Telia to Cogent via Verizon for 12 hours, these routes then disappear, partitioning the Internet for some single homed customers behind Cogent and Telia (they could not reach one another)
• On 28/03/2008, peering link is restored and the Internet is once again whole!

Renesys's Analysis
The following statements are taken (almost) entirely from Renesys.

Why did this happen?

• The peering disputes (Telia-Cogent) tend to be about peering ratios; i.e. one party is carrying the other's traffic longer distances
• Did Cogent view Telia as in a weaker position? (How many European customers want to reach Cogent hosted content?)

Facts after peering link has been restored:

• Number of routes advertised to Telia is increased; Telia reaches 2934 more prefixes via Cogent
• Number of routes advertised to Cogent is decreased; Cogent reaches 635 fewer prefixes via Telia

"This was the result of reengineering by both parties. It surely seems like this was about traffic ratios"

Lessons learned

• Being a (near) Tier-1 is not easy; you depend on everyone else in the cartel and you will be punished if you are perceived to be in a position of weakness
• Peering relationships are tricky; depend on both objective measures and perceptions; disputes can take a long time to resolve; the only driver is market pressure
• Being single-homed is dangerous!

Other notes

• An increasing number of businesses and organizations are now multi-homed
• Many more peering relationships, lessening the impact of any one loss
• True global Tier-1s would probably never consider depeering; impact would be wide ranging; it would invite government involvement
• Tier-1s are becoming less relevant, as smaller players peer around them (This shall be the reasons, why Tier-1s have always to defend their positions with expanding their coverage and acquire new customers !!)

Credits: Renesys Corp. (Martin A. Brown; Alin Popescu; Earl Zmijewski)

Jakarta (IP) Bandwidth Price

I followed some traces left by Pak Meneer and was delivered to his records at Opisboy. This is claimed to be an "Informal ISP Technical ( dir-pji-like :p )" forum. Comments ? :-D Well, for non-local readers, "dir-pji" could be understood as "indonesian ISP Executives/VIPs" or as the forum of ISPs' owners/high-rank officials. It is "dir-pji-like" ? Personally, I would say "better-than-dir-pji" :-D Reasons? he..he..he. Don't take it seriously or even start thinking about that!! :-D

Back to bandwidth. According to Pak Meneer, the condition is as follows:

• Jakarta - Singapore connection costs around USD 150 th./month (STM-1 closed circuit, clear channel); or
• Jakarta - Singapore connection costs USD 1.5 mn/10 years (STM-1 closed circuit, clear channel, IRU scheme)
• IP transit in SG costs around USD 60 - 100 /Mbps/month

From this numbers, he estimates that the cost of (budget necessary for) IP transit at Jakarta would be approx. USD 1,068/Mbps/month for normal scheme, or approx. USD 181/Mbps/month for IRU scheme.

This is more-or-less justified by several recent media reports (quoted information given by Indosat/Fadzri Santosa, APJII/Sylvia or others) which give the figure of around USD 1,000 /Mbps/month. Sylvia and Fadzri might talk on different contexts e.g. Fadzri mentioned closed-circuit leasing bandwidth and Sylvia talked about IP transit. But nevertheless, if we look at the numbers given by Meneer Tutut, cost component for transit doesn't contribute significantly. If Indosat plays in this price level, then I guess, Matrix and Moratel/XL would offer a bit cheaper price.

Regarding IRU scheme, I know that Matrix offers this scheme. But I doubt that Indosat does (will do)?? So, the information above may specific to Matrix. Right pak ?? :-D Some friends of mine have told me that Matrix has done a very good "marketing move". They may soon increase their currently-almost-full-booked 60 Gbps capacity. :-D However this might a spoiler, not official!

Last issue, many people use USD/Mbps metric for bandwidth price. I would personally prefer USD/X-port metric e.g. USD 2,000/E1-port rather than USD 1,000/Mbps. Why? Simply, because not all line bandwidth are useable for user data. If we are talking about STM-1 line for instance, if it is channelized, its useable bandwidth is only 126 Mbps. If it is unchannelized, by using PoS for instance, its useable bandwidth might be in the range of 130 - 140+ Mbps.

That's all. Sorry this time without pics. :-D

Jakarta Metro Backbone

It seems that everybody in Jakarta want to have his own metro :-) Currently many companies have built their metro infrastructure to support their business, particularly in big cities such as the Greater Jakarta, Surabaya, Medan, Bandung etc. These include :

• well known telcos (TELKOM/TSEL, ISAT, XL, other?)
• successfull ISPs (BizNET,CBN, other?)
• VSAT network operators (CSM, Lintasarta, other ?); and
• backbone operators (ICON+,PowerTel,Napinfo,Moratel, other ?)

Recently, PT. iForte Solusi Infotek announced, that they have finished the construction of their fiber optic network which is using (and installed simultaneously with the development of) the TransJakarta busway infrastructure (map see below). They currently offer connection/access services called M-Wifo which are basically WiFi (802.11n compatible) access and fiber bridging. To my guess, it might be a RoF (Radio over Fiber) system.

Other related public information about iForte:

• the fiber optic backbone spans/follows 10 TransJakarta corridors
• CAPEX : USD 10 mn.
• Length : approx. 250 km
• RFS : June 2008
• Services : fiber optic connections, WiFi mesh 802.11n connections
• Projection : +/- 2,000 cust. (connection point) each year
• VIPs: Peter Djatmiko, Andrie Tjioe
  

Map of TransJakarta Busway, Courtesy of Maximilian Doerrbecker (Chumwa), taken from Wikipedia

How Much Bandwidth do Indonesian ISPs Have ?

Though there are many Internet and NAP service providers running their business since several years ago, "big Internet pipes" are only owned by the (former-telco-only) giants: Indosat, Telkom, Excelcomindo, and Telkomsel. They may have a total international IP bandwidth more than 1 Gbps. Unfortunately, as usual :-D I don't have complete, official, clear, and confirmable numbers for that. Maybe next time, I would come with a simple calculation using several rough statistics found in the media and balanced with some cases (personal experiences).

As can be seen in the mindmap below, Telkom shall have international IP connection (transit + peering) at least 2 Gbps; Telkomsel at least 1 Gbps (It is for me still a question, whether Telkomsel shares the bandwidth from its parent PT. Telkom ??); Indosat at least 2 Gbps (I found different numbers for year 2006/2007). Somebody will to advise ??

National and Regional RENs

Actually since a long time ago I want to write threads about RENs (Research and Education Networks). Why? Simply because they were the origin of the current Internet and they still have a major role in creating advances in networking technologies. This time I would not start with an indonesian case (if INHERENT or Jardiknas can be categorized as a National REN/NREN), but with european cases.

BWiN, GWiN and SurfNET

GWiN Connectivity and Traffic Statistics

Each country in Europe has one or more NRENs. These are connected at a higher level with regional european RENs. Some of you might already heard about TEN-155 or currently GEANT/GEANT2; these are RENs at european level. At country level, Germany for instance has formerly BWiN (Breitband Wissenschaftnetz), GWiN (Gigabit Wissenschaftnetz) and currenly XWiN (X Wissenschaftnetz :-)). Other examples: France has RENATER, Holland has SurfNET, UK has Janet and UKnet?. During my dissertation several years ago, I used intensively these networks to measure, calculate and simulate performance of my approaches and algorithms.

These RENs are active and very dynamic since they are intended to fascilitate research on advance networking technologies and their applications. We may take a look the case of the RE networking (r)evolution in Germany:

First Generation (1989/1990) --> X.25-WiN

• 64 kbps in core; 2.5 - 64 kbps access
• X.25 technology
• 1990 upgraded to 2 Mbps (core and access)

Second Generation (1995-1999) --> BWiN

• 155 Mbps in core; 4.8 kbps - 155 Mbps access
• ATM technology
  

Third Generation (2000-2005) --> GWiN

• 10 Gbps in core; 128 kbps - 10 Gbps access
• PoS technology
  

Fourth Generation (2006-) --> XWiN

• heterogeneous technologies based on recent advances in optical transmission/networking

At a glance, from my point of view, XWiN takes the NLR (National Lambda Rail) model to some extent.

International IP Bandwidth

Currently I have only some official information specific to Indosat. This was presented by Indosat several times ago (Apricot 2006 Bali ??). However, it belongs to the "biggest" pipes owned by indonesian ISPs.

Indosat's IP Transit Bandwidth:

• Batam - Singapore --> 465 Mbps
• Jakarta - Singapore --> 930 Mbps
• Jakarta - Hongkong --> 775 Mbps
• Jakarta - USA --> 620 Mbps
  

• Total --> +/- 3 Gbps
  

Indosat's International Peering Bandwidth:

• Total --> +/- 500 Mbps
  

Can Broadband Spread to the Masses?

In this post, I just want to archieve some facts and opinions given by Arjun Trivedi (NSN-Nokia Siemens Networks) in the first Indonesian Broadband Summit held last April in Jakarta. The title above is taken from Arjun Trivedi's slides, since to my opinion, that title describes the most important message from his presentation. (I was not attending this event and just found his presentation material through a link referred from Mastel website). But don't forget to balance the question with "Where are the Masses?" :-)

Some facts/statistics as of end 2007 (source UN, Ovum, ITU, BIS Sharapnel, Operators, NSN Market Intelligence):

• Population 233 mio.
• Households 59 mio.
• Internet users 31 mio.
• Dial-up access 0.7 mio. subs/month
• Broadband subscription 0.3 mio.
• HSDPA users 0.1 mio.
• PC/laptop households 7mio.
• DSL subscribers 72%, Cable subscriber 25%, other 3%
• Broadband CAGR Indonesia 115% in last 3 years (2005-2007)
  
  

Regions served by broadband according to NSN

Opinions:

• Broadband is still in early stage of development with sporadic coverage (see above)
  
• Broaband access can open up a nation's economy once we overcome challenges in affordability and connectivity: increased competitiveness, better education systems & working environment, savings of operational costs
• Key drivers for broadband penetration : acceptance of Internet, affordable packages, variety of technology choices & availability, government support for ict development
• DSL is still preferred form of broadband access globally; cable modem sparks demand with TV offerings; other technologies such as FTTx, WiMAX are in the pipeline for introduction
  

Profile of indonesian broadband subscribers (NSN surveys 12/2007-2/2008; Jakarta, Bandung, Surabaya, Medan)

• Typical user profile: highly educated people in their early 20's
• Common user habbit: daily home-working, email, browsing & chatting (PC/laptop)
• Perception of broadband: faster internet access, simultaneous use of telephone and internet
• Most familiar broadband service technology: DSL, Cable
• Preferred billing system: post-paid (more specific ? - may be "flat" but "cheap" :-))
  
• Preferred method of payments: ATM, authorized dealers
• Main influence in selecting broadband access: recommendation from friends
• Barriers to have broadband access: price and service availability

Some other important issues for broadband acceptance:

• Good reputation
• Extensive network coverage; connection quality; stable speed
  
• Ease of payment, accurate billing, easy understanding of bill statement
• Price affordability, fast installation, good value for money
  

Case "Broadband in South Korea" :

• Over the past two decades the SK government and its privatesector have built one of the world most advance IT infrastructures
• 76% of Koreans over age of 6 use Internet with e-government, e-commerce and e-education being popular applications
• In 2006, korean IT industry grew approx. 4x more than the non IT industry, and its portion of total GDP is over 16% while telecoms services account for over 5%
• Government established "u-KOREA Master Plan" and part of this is the establishment of "u-Infrastructure" or a 'Broadband convergence Network' (BcN)
• BcN aims to provide broadband multimedia services to 10 mio. households (approx. 55% households) with wired services and 10 mio. subs (approx. 20% of the population) with wireless services by the year 2010 and to achieve telecom/broadcasting equipment production volume of about USD 73 bill. as well as export volume of USD 20.1 bill
  

Indonesian DSL Market

The title might be broader than I could write here right now. However, as usual, the thread will be continued later if I get some new interesting information.

Though several small ISPs do offer DSL access, PT. Telkom is surely the market leader with 97+% market share (of the number of customers). You may see the number of customers as of Q2/2008 in the picture below. There are totally around 400 th. subscribers (we might compare with the number/statistics mentioned by Arjun Trivedi - however it is for Q4/2007). More than the half of it (200+ thousands) resides in Java.

Other rough statistics given by PT. Telkom:

• Customer acquisition rate: Q1/2008 --> 567 cust/day; May/2008 --> 1081 cust/day
• ABPU (Average Bill Per User - any better terms ?? ): IDR 314 th./month
  

Indonesia's (ICT) Statistics

In this post, I just want to archive some statistics, in particular those with regards to telecommunication / ICT development in Indonesia. If it is not explicitly mentioned, the values are valid as of Q1/2008.

ICT statistics (KOMINFO):

• Wireline telephone subscribers : 8,758,096
• FWA (limited mobility) subscribers : 9,069,827
• Mobile subscribers : 81,834,590
• Internet subscribers : 2,543,649
• Internet users : approx. 10 times of the number of Internet subscribers
• Fixed broadband subscribers/users : approx. 241,000 (ADSL, Cable, WiFi Hotspot -- actually it shall be more!)
• Mobile broadband subscribers (3G, 3.5G, ...G :-)): 5,500,000 (??)
• Density: PSTN-->3.89%; FWA-->4.03%; Mobile-->36.39%; Total-->44.31%

• 182 ISPs (only 43 active -- number of subs. as mentioned above)
• Computer penetration : 0.01-0.05% (less than 5 PCs / 100 households)
• Thousands of public Internet kiosks (Warnet)
  

Players/Telco : (KOMINFO)

• PSTN : Telkom, Indosat, Batam-Bintan Telekomunikasi
• FWA : Telkom, Indosat, Bakrie Telecom
• National long distance (SLJJ) : Telkom, Indosat
• International long distance (SLI) : Telkom, Indosat, Bakrie

• GSM & 3G : Telkomsel, Indosat, Excelcomindo, HCPT, NTS
• CDMA : Sampoerna (STI), Mobile-8, Smart Telecom (Sinar Mas)
  

General country statistics : (KOMINFO/www.datastatistik-indonesia.com)

• Area : 1,905,11 sq. km
• Population : 224,904,900
• GDP : USD 467 billion (USD 2,053 each)
  

P2P Internet Traffic : The Elephant

Since several years ago, peer-to-peer issues have been a central topic on IP networking. This has been pictured in the program of some top networking conferences e.g. INFOCOM, SIGCOMM, ITC, GLOBECOM etc. What actually makes P2P issues so important?

One possible answer can be seen in the graph at the right. According to Telegeograhpy, as of 2007 54% of the total Internet traffic is originated from P2P applications which are in the forms of video (61%), audio (11%) and other formats (27%).

Some other facts (taken from Telegeography):

• P2P traffic accounts for nearly 60% of Internet traffic (Sandvine)
• The share of P2P traffic varies from 10% to 70% depending on a carrier's customer base and geograpic coverage (TG's interviews with carriers)
• When TPB torrent network was taken down in Sweden, TeliaSonera's Internet traffic declined by 60%!
• P2P accounted for 50% of upstream capacity on Japanese networks in 04/2006 (Japanese Ministry of Internal Affairs and Communications)

Applications that may or may not impact long-haul networks:

• YouTube serves 70 mn low-res videos daily!
• Online Gaming: World of Warcraft reported 6.5 mn subscribers in early 2006
• Video Calling: Microsoft claimed that users of Microsoft Messenger engaged in 1.1 bn minutes of video calling during 01/2006 (how about Yahoo?)
• Sporting events which routinely streamed online (World Cup, major league baseball etc.)
• Grid Computing : Teragrid (10 Gbps connections)
• In 06/2006 more than 106.5 mn people (3 of 5 US Internet users) streamed or downloaded video (in total, nearly 7.2 bn videos were streamed or downloaded - comScore 09/2006)

Intra-Asia IP Traffic

This post is closely related to the previous one, except that here we discuss about Internet traffic not about bandwidth capacity. I found an interesting statement given by a REACH official:

"In 1998, 100% of IP traffic between countries in the APAC was transmitted via the US. In 2004 it was around 30% (estimated)."

I do not know what criterion is used by REACH to come to those figures. But pesonally, I feel it is highly possible to be true. In around 1995, I remembered that when I called "traceroute" from my university to any local ISP, it gave me several US-hops! At that time I was dreaming a local version of NSFNET in this country (yes sure, later I knew that this is the timeframe where NSFNET was decommisioned!).

Many people at that time only concerned on establishing connections to the center of the Net i.e. the US and did not care about anything else. But slowly the "message" is spread and more and more people understood the idea behind this new thing. ISPs were making both bilateral and public peering(*). The govt has updated the regulation for encouraging local connections. Is it enough to suppress the inefficiency? Maybe! But I believe that right now intra-Asia IP traffic that travel first to the US is still in the figure of 10% or even more.


(*) for local readers, please note that peering != "piring" ! :-D

Intercontinental Internet Bandwidth

Recently, Telegeography has published some information and facts for the telecommunication's world in the last year. One of the facts is related to the international IP bandwidth as given in the picture (at the top).

Surely, except the numeric values, it might for many people not a surprise : North America is the center of todays' Internet. This actually has happened already for a long time, since the born of the Net :-D. As of 2007 the total aggregate bandwith capacity to/from North America is around 2.7 Tbps, a half of which is used to connect to Europe. Asian countries have an aggregate Internet bandwidth of around 0.9 Tbps, which is still below that owned by the european countries (1.7 Tbps). The capacity for Non-US/CA connections is only 0.2 Tbps. What does it mean ? The Internet would practically not work without US and Canada :-D ! This might also imply that US and Canada host the majority of the current Internet content!

But the figure might change in the next years. Though the increase of trans-atlantic IP bandwidth is still dominant compared to trans-pacific or other regions (see the picture at the bottom), the role of asian countries could in the future become more significant. The reason is simple: the asian countries are densely populated and have many potential Internet users/players!

Domestic Bandwidth Price

According to APJII, the Indonesian ISP Association, this year the price of international Internet bandwidth in Indonesia tends to decrease (reported by detik.com on 29/January/2008). Last year the price was in the range of USD 2.200 - USD 2.500 for each Mbps. But since early 2008 it has been reduced to around USD 1.800 for fiber access and around USD 1.300 to USD 1.500 for satellite links.

Cheap enough? I don't think so. I guess the price in India or Malaysia is (much) cheaper. I don't have any figure right now. But last year, in an informal discussion with some key persons from a malaysian Internet player, the price ratio MY:ID is around 1:4. Right? Malaysian readers, please CMIIW ! Surely, I'm not talking about TM :-D. Otherwise, if you have access to the paid version of Telegeography databases, you can easily compare the prices and drop here a line :-D

Good news ? Maybe. I see that the most indonesian ISPs are moderate and innovate very slowly (if any). Such a price effect may or may not touch end-users. And if it indeed reaches the customers, it might take several months to propagate, or even worse.