The MEA DNS Wars (1998 - Present) - Episode II - Part 1

Disclaimer & Opening Commentary

The term “DNS Wars” is not mine. Originally it referred to a rather weird phase (the “Dark Ages”, if you will) of the global Internet governance, eventually giving way to a less US-centric and more inclusive governance framework, under the rule of the ICANN (The Internet Corporation for Assigned Names and Numbers). It is used here neither as a throwback to that “wild west” era nor as a cheeky “clickbait”, but as a logical way to elevate the discussion upwards onto the next higher layer of the Telco/Internet hierarchy; moving it from the physical infrastructure layer covered in the MEA Cable Wars 3-parts article to the IP layer on top. This article could have been titled “MEA IP Wars”, but I didn’t feel it would have the same ring to it.

As used in this article, this term – therefore – will cover the IP/IPT services landscape across MEA in addition to looking at MEA DNS data and their implications. The discussion emphasis will be on how Telco operators managed their positions competitively. It will also - as was done in the MEA Cable Wars article - try to build a national/regional context to those commercial & operational activities so that an understanding develops as to who are the leading markets/economies across MEA in this arena.

As you must've surely noticed from the article's title above, the MEA DNS Wars started in 1998, 2 years later than the MEA Cable Wars started. This is consistent with the development paths of both the MEA Telco & MEA Internet markets. Unlike the MEA Cable Wars, though, the MEA DNS Wars have NOT ended yet, so there are still some interesting battles ahead across the region.

What was at stake?

Similar to the MEA Cable Wars, the sought-after prize in the MEA DNS Wars was the so-called “Regional Hub” status. The objective – both at the Telco operator level or the national level - was to become the gateway through which other Telcos/countries got access to the Internet. The role of “regional Internet traffic aggregator” quickly emerged as the latest coveted proverbial feather in a Telco’s cap, and specific countries were keener than others on achieving that objective.

A word on terminology is needed to thwart potential confusion. The term “Internet Exchange” or IX/IXP is fairly well-developed in the US & European contexts and implies certain guarantees about AS/Network Reachability, Public/Private Peering, Security, Service Levels, etc. In MEA, however, Telcos (in the absence of a clear regulatory policy mandate) found it way easier & more marketable to label their wholesale Internet Access services (IP Transit) as an Internet Exchange (IX/IXP). Proper “Carrier Neutral” IXs/IXPs only started emerging in MEA over the past 3-4 years only.

Battlefield Dynamics

Unlike in the submarine cables space, where several MEA Telcos were primary investors in international/regional cables, and hence enjoyed competitive cost advantages when competing regionally for connectivity services bids, the region’s Telcos were at an international disadvantage in the Internet space. This was a reflection of the asymmetry in IP traffic volumes that originated & terminated on MEA networks.

In other words, even the biggest MEA Telco was effectively a Tier 2 IPT provider in the Internet’s hierarchy, as they – at best – bought wholesale Internet access from a Tier 1 global provider. Still, the model was a lucrative proposition for those MEA Telcos aspiring to become a regional aggregator through 2 distinct value drivers:

• Selling the access circuit cable capacity so that their regional customers can reach the aggregator’s network.
• Oversubscribing their purchased IP capacity multiple times when they sold IP Transit to the tier 3s & tier 4s regional customers.

The above service configuration was super-profitable commercially. Telcos made money on the IRU pipe sale to their customer (with annual O&M payments over a 15-year period) on cabled they were part-owners of, then again on the monthly/annual lease for the IPT port (which was not really dedicated, allowing the Telco’s to overload their routers for maximum gain.

Of course, not every Telco with global network reach & substantial cable capacity inventory played in that market or showed interest in becoming a regional MEA IP aggregator. As times passed by and the wave of CDNs/OTTs; and later Cloud Service Providers/IaaS providers hit MEA, those Telcos were facing more operational & commercial pressure to meet customer demand for reliable & fast Internet capacity.

The choice of how to engage in this regional competitive market came down to one of only 2 alternatives:

 A.  Buy wholesale IP in Europe/USA/Asia Pacific (at the STM-1, STM-4, or 10G IP port speed levels), long-haul that port connection back to the Telco’s POP back at their home network/country, then sell their own port connections regionally. The Telco had end-to-end control and – at least for the first 12 years or so – regional customer expectations were low (just give me wholesale Internet access) and so were not difficult to meet.

B.  Host a global Tier 1 or Tier 2 IPT provider’s POP (with significantly-large global AS reachability) on the Telco’s network (inside their home country), and sell access circuits to that node, plus a share of the revenue on the IPT port’s monthly/annual charge. This option became necessary as regional MEA Telco customers were no longer satisfied with the high latency for the earlier long-hauled service configuration.

Eventually, global Internet Exchanges (IXs) started moving into the region. These were carrier-neutral and in many cases were set up as "not-for-profit" organizations, so open interworking & peering were primary mandates, that carriers had to adjust to.

Leaders & Followers

The below chart (credit: TeleGeography's Kate Reilly, originally published here https://www2.telegeography.com/capacity-middle-east-2019), shows the total Internet bandwidth distribution across MEA. This combines the national/organic IP bandwidth for each country with the regional/inorganic bandwidth (which is sold to regional customers and aggregated at each location on the map towards the upstream IP backbone provider).

A few high-level comments can be made here:

1) Only a handful of locations managed to aggregate enough IP bandwidth to be regionally relevant. Those leader locations are the natural candidates to become the MEA region's "Cloud On-ramps" and hosts.

2) It is an interesting exercise to superimpose the MEA regional cable map on top of this map, to get an understanding of how well some carriers added value to (or “leveraged”) their regional fiber networks to build critical mass on the IP layer. As you will find, not every location made strategic use of their underlying infrastructure.

3) If you are not a leader, you become a follower. This dependency relation can be manageable/acceptable within the Internet access space but becomes trickier as the region moves towards the cloud-centric computing model, as then the follower country to face the possibility of having mission-critical applications for its own top banks, businesses, and possibly its own government homed in another country.

4) Overall, the total IP bandwidth serving the MEA region remains insignificant on the global scale.

Do you want to see how rapidly the regional Internet bandwidth picture changed? Consider this 2015 snapshot. Nothing more needs to be said, really.

One final "macro" comment is needed, and that's on wholesale IP Transit pricing in MEA. Without exception, this remains one of the most expensive regions to purchase IP Transit in. It is not uncommon to pay 10 to 15 times more per Mbps for a local IP Transit 10G (Gig-E) port compared with developed markets in Europe (London, Frankfurt, etc.). As the region moves from a Telco-centric to a Cloud-centric environment (whether everyone likes that transition or not), the price of IP Transit will be a big development hurdle, which will inevitably need regulatory intervention to resolve.

End of Part 1

Part 2 (next week): Regional Internet market data, E-commerce considerations, and national considerations.