Pakistan Bandwidth Appetite

I had been meeting a lot of fellow friends in the industry in person lately for a number of reasons. One interesting number that I came out with was that of the total transit Internet bandwidth that Pakistan is currently subscribing to. Between PTCL, Transworld, Multinet (the companies) and SMW3, SMW4 and TW1 (the three cables -and a couple of transponders worth of satellite back up bandwidth) we are talking about an aggregated 2.5 Gigs of IP transit bandwidth (give and take an STM-1).

Not only the figure is respectable, the growth trend that my contacts described is excellent. Drivers for increased consumptions, as we discussed and nodded at, included new DSL services, VoIP traffics and the popularity of bandwidth hogging Internet services like Youtube.

A number of other potential bandwidth-suckers that are waiting around the corner include mass distribution alliances that are expected to talk place between bulk bandwidth suppliers mentioned above and the new distribution channels like Niyatel’s (and other similar outfits in Central Pakistan).

This trend is impressive. Additionally, it also presses the managements of the local industry to come forward and think broadly out of the box for some badly needed but missing standard industry practices like local peering and Internet Exchanges and local Data Centers. If the growth trends continue, it can soon become a major foreign exchange eater and the only way we can reign this in is through localization of traffic within Pakistan.

Advertisements

Pakistan-India Fiber Linkup

There are more news about the Pakistan-India fiber linkup via the Wagah border (Lahore/Amritsar). Essentially, the project is inching, not leaping, further. As we discussed back in early 2006, India is said to be reluctant about the equality of the utility of the optical fiber. From the The News story:

It’s not a big issue, but it has to be resolved first before it gets operational,’ said a source close to the development. ‘There are more bottlenecks on the Indian side compared with Pakistan, as India believes the project would benefit Pakistan more.’

An optical fiber link is also being established between Pakistan and Iran.

Its The Network, Stupid!

PTCL has recently updated its tariffs for Internet bandwidth that most of the Internet Service Providers in Pakistan subscribe to. The new rates reflect a 37% reduction. A transit Internet capacity of and E1  now costs $1,000 per month ($500/mb). The new rate table is:

S. No.

Speed

Charges (US$)
Per Month
At Karachi

Charges US$)
Per Month
other than Karachi

1

2 Mbps

1,000

1,200

2

45 Mbps

15,000

18,000

3

155 Mbps

30,000

36,000

Due to obvious reasons, this blog had been following bandwidth’s price reduction in Pakistan. We have seen the price of 1 E1 coming down from  $6,000 per E1 to the current $1,000 in around four years.

Transit bandwidth costs getting lower means better business costs for ICT enterprises. However, for ISPs whose business in Pakistan was thriving on the ‘scarcity and preciousness of bandwidth’ will get effected fundamentally. Until now, ‘building’ out the access network had not be a top concern of most of the Internet service providers in Pakistan. Selling any precious commodity – be it bandwidth or saffron, typically becomes a business where needy customers will come to you and you can make your margins on the the transaction.

With bandwidth prices getting more aligned with rates of the same elsewhere,the question of ‘owning your users’ (or, in other words, building your network) would gain importances. Typical ISPs today have a number of wireless or optical fiber links from their prime customers coming from various location (in each big city) to (typically just one) their facility which has bulk bandwidth  provisioned. The higher costs of bandwidth in Pakistan had traditionally allowed ISPs to hide the high costs of such isolated last miles within the total price of the deal and get away with this inefficiency of (not building an access network of) their own.

Lower bandwidth rates, widely advertised and discussed, will now prompt the customers to question the selling cost of the bandwidth. This will quickly point to the fact that in a poorly provisioned general telecom environment (bad copper, lack of fiber etc.) the absence of ISPs own access network is contributing to the total cost of the link.

ISPs will now be forced more than ever to build their networks in major cities to make sure that transit bandwidth still remains the only major element in their offered prices to their customers. Current practice of high price wireless and fiber links (all being pulled to a central location) must go away.

It is also important to note that PTCL has now, for the first time, started making the price differentials between Karachi and upcountry obvious.  The cost of carrying the bandwidth from Karachi (where it typically enters Pakistan via the submarine cables) to up country has always been there but has always been absorbed in the model and the industry has been presented with a uniform rate across Pakistan. The new private ownership of PTCL cannot afford such political luxuries and is making the costs obvious in its tariffs.

This should also send a ‘think-again wave’ to TWA and other operators who were planning to offer transit Internet bandwidth on uniform rates  across Pakistan. TWA was reportedly banking on the transit capacities across Pakistan that it is taking from Mobilink’s under-construction long haul optical fiber network. Unless the capacity comes free of costs, the price has to be paid either by the customers or subsidized by TWA which appears to be unlikely.

Summarizing, booming bandwidth consumption and lower bulk bandwidth costs are pushing up the pipe sizes for Internet across Pakistan which makes networks costs considerable. Considerable network costs mean that they do not get subsidized anymore and for end ISPs, owning their network will start making more and more business sense. Is the lot ready?

Competition in Domestic Capacity Sector

In an effort to enter other segments of telecommunication services beyond cellular voice services, Mobilink is now actively marketing capacity on its national optical fiber network to potential customers. The company has established  a separate business division to cater to this vision. Primary targets are alternate careers and long distance voice operators that are currently subject to the monopoly of PTCL in this segment.  Blue chip customers that might have sizable requirements for domestic data traffic are also being targeted to some extent.  Typical discount rates that are being offered to existing PTCL customers are in the bracket of 20% to 40% depending on the business size and the customer(s) market standing.

Last mile issues – the network segment from the potential customer to the service provider – is also being handled very aggressively. For bigger customers where ROIs are shorter and reasonable, the company is reportedly willing to build microwave links to shift the customer from its current provider (PTCL) to itself.

Customers have shown a mixed response to the sudden availability of competition in this expensive segment of telecommunication services. Last mile issues mentioned above are not always sorted out for smaller customers who are required to build their own network up till the new providers point of presence. Unavailability of dedicated service organization within the new alternative long distance capacity providers is also a major turn-off point for the potential customers majority of whom are not ready to buy the argument of these company which revolves around the myth that the service will be error-free because we (the providers themselves) are running out network on it.

A major driver toward these new service providers could be the need of businesses to diversify their long-haul communication infrastructure by going to two different service providers’ networks. However, in the absence of services packaged specially for backup purposes, the high costs will keep a good number of customers away from the new providers.

Pakistan-Iran Fiber Link

Pakistan and Iran started talks on a possible terrestrial optical fiber connectivity between the two countries back in December 2005 have now signed a Memorandum of Understanding for the same. The deal between Iran’s Telecommunication Infrastructure Company and Pakistan’s state-owned NTC. Capacity of the fiber would initially be 64 STM1 links.

Once realized (at a reported cost of 1 million dollars), the link will be the forth major optical  fiber connection of the country to the outside world. Currently, SMW3, SMW4 and Transworld’s TW1 are the only three physical optical fiber cable systems (beside some satellite based connectivities in Karachi and Islamabad) that connects Pakistan to the rest of the globe. Talks for a link via India are on since since December 2004 but the deal has not yet materialized and no circuits have yet been fired up.

Paktel’s Continued Transformation

Cellular News is reporting that Paktel’s stakes were acquired by China Mobile Communication Corporation which is the parent company of China Mobile. China Mobile is a listed company and the move is intended as the investors were cautious on the profitability of the venture in Pakistan due to fierce market competition.

Insiders are reporting that the new company, now renamed as CMPak has started a fresh wave of hiring employees.  The company is also engaged in talks with a long-haul optical network with Malaysian roots to acquire dark fiber pair on long term lease for its nationwide traffic requirement.

PTCL Launches Own DSL Operations

PTCL has launched the anticipated DSL services under a confused brand name today. Apparently branded as ‘Broadband Pakistan’, the service seems to be off to a hasty start.

DSL was one area PTCL has largely left to O&M operators (who use the copper infrastructure of PTCL for their DSL services) and to its practically estranged subsidiary of Paknet. The ‘tariff’ page of Paknet has gone 404 probably after PTCL itself took up the DSL business which it announced today.

The sign-up page for new subscribers is broke. The form asks for a city but does not give any option to provide one. The phone number provided on the ‘customer services’ (800-810-999666) does not work and calling it gives you a recording for an erroneous dialing from the very network of PTCL. The ‘online Help’ link is un-clickable.  Presumably a placeholder entry by the webmasters that carried out to the live website. PTCL is looking out for big guns but it seems that for now, they need some good web development team that can deliver a coherent and functional public interface for them.

Keeping the ‘teething problems’ aside, the plans are interesting and suggestive of the future trends the DSL market is going to follow. The base plan starts at 256 kbps so there is no 128 kbps option altogether. No security deposit, no start up fee, free CPE and unified billing.

Readers are requested to share their experience of the service as it becomes available for the rest of us.