Broadband Penetration - MoITT, USF @ Work

Universal Service Fund (USF) is the company formed to make use of the USF money that PTA has been generating out of the booming telecoms market of Pakistan. So far, USF has worked towards using its funds for the spread of voice services in the under-served markets of Pakistan. Of late, Ministry of Information Technology & Telecommunication has intended to guide USF to do the same towards increasing broadband penetration too.

USF, after some initial work, has concluded that there are no particular areas that could be defined as ‘under-served’ in terms of Pakistan and rather the entire Pakistan is under-served. USF has now asked MoITT to pass a ‘determination’ towards the same fact allowing USF to utilize the funds anywhere and everywhere in Pakistan.

MoITT has published a 39 page study document on the web which seeks to establish this fact (that the entire Pakistan is under-served in broadband services). A consultation session was held in Islamabad yesterday to discuss this matter with the industry. The proceedings and details of the session are still to come out but here are my initial takes on the document and its contents:

The  major conclusion points of the documents are:

• Pakistan’s broadband penetration is very low
• Currently there are around 100K Broadband subscribers which need to be taken to 1.6 million by 2010 (1% of population)
• This low penetration is earning bad scores for us under the WSIS measuring criteria & there is a strong need to improve the same
• Three approaches have been suggested for the GoP’s intervention in this ‘dismal’ state of broadband affairs:
  • No intervention - leave it to market; slow broadband growth expected
  • Bundle with Basic Services - only rural areas will benefit; existing broadband provides will loose
  • Tackle issue with a new format - dedicated efforts are expected to yield better results; divided in various phases

The document assumes or maintains that fixed broadband is a dwindling trend and wireless broadband will finally prevail (page 23). While this is true for the last mile domain, the infrastructure is ALWAYS wired (read fiber). The guys at the MoITT need to be pointed to this omission in consideration. Pakistan need to have a good wired infrastructure before we can decide which of the two last miles options (wired or wireless) is good for us.

The study also repeatedly mentions the similarity between low tele-density and low broadband penetration. However, the applications/demand side difference between the two (voice and data) is repeatedly ignored. While it is true that the gap between 2.7% tele-density (from where our telecoms boom started off) and current 50% tele-density was one of the reasons for the boom, it was the application (voice) that was ready to exploit this gap. In the case of broadband, a similar gap exists and this gap is what the study is considering as an opportunity. However, as obvious, the difference between our last success (in cellular voice) and current challenge is that of application - do we have compelling applications that will drive the growth that can ride this gap?

The document also does not considers demand creation at all. While supply end enhancements (by way of USF subsidies towards network deployments etc) are more than welcome, a significant portion of the efforts must go towards demand creation activities. Mandatory use of electronic facilities in the business circles, tax cuts for ISPs interconnecting with each other, financial benefits to private TV channels to host streaming servers inside Pakistan, creation of public/open Internet Exchanges etc are all example of such efforts.

IXP in Pakistan

PTA is soliciting proposals for Consultancy Services on the issue of establishment of local Internet Exchange Points in Pakistan.

The last date of submission of such proposal is around the end of Feb 2008. Let us hope PTA gets good consultants to get them going in the right direction and speed.

In this relation, here is an interesting presentation on IXP by Guarab Raj of SANOG and PCH fame.

The Case for an IX

This is about establishing local clearing of Internet traffic. A concept so basic to the Internet that unfortunately remains grossly neglected (or not acted upon due to misplaced fears and priorities by the parties involved).

Before you shouts the word PIE, please note that local clearing should generally have two traits

• it must happen inside the country
• avoiding (or minimizing) expensive for-ex costing international links
• at next-to-free bandwidth rates - using the SKA (sender keeps all) or cheaper local bandwidth models

Focus Karachi (2007) for example:
Three known names to average broadband/dialup users: Worldcall, Cybernet & Multinet. Despite some efforts of optimized routing, the majority of the traffic between these well-known ISPs is exchanged outside Pakistan. The first one is on Transworld (now becoming a mini PIE on its own due to a good amount of ISP customers on its network), the second is on PIE and the third on its own IPLC towards KL in Malaysia.

Short Term Problems:

• Uniform (high) Internet prices across the industry
• Unnecessary 20% to 35% for-ex spending on Int’l circuits that could be sold to the right customer at a premium
• No cheap ‘local bandwidth’ available to users
• No incentive for ‘remaining local’ to users or content publishers
• Network outages beyond Pakistan result in network outages inside Pakistan

Long Term Problems:

• No impetus for establishing local data center for Internet hosting needs (excluding vertical DS such as Banking etc)
• As more of the country embraces Internet, ’short term problems’ identified above will get magnified.
• No mirrors of popular contents - even those willing to place their contents near the Pakistani Internet users are amazed at the absence of local IDCs.
• No real development of Urdu and other regional languages contents on the Internet. BBCUrdu to remain flag bearer of Urdu on Internet!

As time passes and Internet subscription of these players increases, they will have an automatic incentive to interconnect locally between themselves on private funds driven by savings-as-profits targets. This is considered a positive development and we are all for it.

However, the absence of a neutral NAP or IX discourage non-ISPs willing to get benefit from local cheap Internet bandwidth because the bigger players might want to preserve  their respective ‘exchange-locally-charge-internationally’ status.

Neutral NAPs and IX need data centers and meet me rooms so it might sound like a catch22 at first. However it is not. With at least two to three facilities based optical fiber Internet service providers in all three major cities of Pakistan, a feed-the-goose thinking by the bigger players is the need of the day. Grow the base market and the business will grow in turn.

Karachi again: There is one TIV certified data center in Karachi now. It has a decent MeetMe Room. We have fiber providers in Karachi who can a) use it for themselves to come there at the MMR and b) offer dark core solely for this pull-them-to-IX effort to at least one of their sizable competition. And in here lies the catch. Do they have the heart big enough to do that? For towel-manufacturers head, this is a tall order.

Could someone think loudly on the same lines in Lahore? and Islamabad?

Comrades, help us all think aloud!

p.s I have nothing against towel-manufacturing industry. Just using it as a metaphor.

Wateen Bulk IP BW: Sales Calls Begin

Wateen Telecom has started making inquiry calls to potential bulk transit Internet bandwidth customers. While price points were not available with the calling representative of the company, they are planning to offer a single rate for IP bandwidth across Pakistan which will be major impact news.

Whether IP bandwidth would be subject to VoIP service restrictions or not is not known yet. However, unless canabilization is not a problem for Wateen which is also actively pursuing domestic players for nationwide transmission services, most of the tier two players would want to opt for the flat rate Internet across Pakistan and use the medium to save their transmission costs.

Currently, PTCL offers transit Internet bandwidth to Internet Service Providers at a flat rate across Pakistan. However, the same bandwidth could not be used for transporting voice traffic by any player as major VoIP services are disabled on this bandwidth. (PTCL has  traditionally been using a so called ‘Network Security & Surveillance Wing’ to actively detect voice traffic on PTCL provided links to ISP and other enterprise customers).

Backbone Interconnections: Wateen would initially be buying bulk capacity from SMW4 consortium members.

Internet Exchange and Private Peering: Busy in their network roll-outs and customer acquisitions, and want of realized volumes, none of the players (TW, PTCL, Wateen, Mobilink) have so far announced any private peering arrangements. Similarly, the Internet Exchange plan also remains a dream because the country lacks agile nationwide ISPs that own a lot of IP users and hence would potentially benefit from an Internet Exchange.

DIDX Gets Internet Telephony Excellence Award for 2006

DIDX - DID Exchange from Super Technologies- with Rehan Ahmed Allahwala from Karachi (yes he travels a lot but is based in Karachi) as one of its main drivers bags one of the forty three companies that received the Internet Telephony Excellence Award for 2006. The award is organized by TMC, a galaxy of communication related print and on-line publications.

Other companies that bagged the award included Alcatel, Inter-Tel, Inc., Juniper Networks, Netcentrex, Overture Networks, Samsung, Toshiba and Verizon.

Super Technologies have managed to remain ahead of the technological curve specially when compared to local companies. As early as in 1995, the company that was later to become Super Technologies was offering Faxaway services (international fax over email) to the local market. At that time, international calls cost you an arm and a leg and email was still a novelty. From Faxaway to DIDX, the company rolled as many services as the number of alphabets in English. Some of which that I can remember off the hand include Internet telephone line where the users used to get a US number for their soft or hard IP phones, call center solutions, IP PBX etc.

DIDX works on the novel concepts of useless-for-me-but-gem-for-others concept of sharing and exchanging DID (Direct Inward Dialing) numbers between Telecos of the world. Each Telco that operates anywhere in the world has access or ownership of the local E.164 telephone numbers. Using DIDX services, and the magic of VoIP protocols, Telcos can offer their cheap DID resources to others and gain DID resources on other networks for their customers use.

DIDX assumes cheap, resilient Internet available everywhere where its services are used - a situation that cannot be taken for granted in contemporary regulatory environment of Pakistan. However, despite the fact, DIDX is picking up in Pakistan. A number of Telcos are opting for DIDX services and offering a myriad of telephone numbers atop their local loop services in Pakistan.

PTA Keeps on Lipservicing Broadband Cause

PTA has yet again paid lipservice to the cause of wide broadband penetration in Pakistan. At a seminar organized on the theme of Wimax by South Asia Middle East North Africa (SAMENA) Telecommunications Council, PTA’s chief repeated the combo of words that could easily be created by anyone associated with the industry.

More clear, practical announcements on the broadband penetration cause would have been far more useful. Yes, PTA has published a consultation paper that aims to bring down the costs of Internet bandwidth in Pakistan but bandwidth tariffs are just one part of the broadband proliferation. Matters relating to IP traffic exchange between service providers, local loop unbundling, application agnostic Internet infrastructure, development of localized contents, IPv6 adoption via incentives and legislation, subsidies on local telehouses in the private sector and many other factors are needed to kick-off a broadband revolution in Pakistan in its true sense.