This article is part two of the first article that looked closely at the subsea cable landscape across Africa. It concluded that nearly all the countries where economic progress and telecom liberalization is evidently gaining momentum are served by two or more subsea cables has created a vibrant landscape for Enterprise Network Services to flourish with a bouquet of connectivity options at competitive price for enterprise customers.
In this context the recent uproar on disruption of ACE subsea cable 10 days back that plunged Mauritania in Internet blackout for 2 days raised awareness on the need to have at least two subsea cables for resiliency. While subsea cables provide the underlying foundation of physical media and high capacity connectivity superhighways, to provide WAN or VPN service for Enterprise Network Connectivity to global and regional enterprise customers that have sites in one or more African countries, network infrastructure of PoP setups leveraging subsea cables is needed that provide MPLS, Ethernet and IP services and combinations thereof. This in turn will pave the way for wider service ecosystem spanning, SDWAN, Cloud, MHS, MSS and UCC to grow. Over the last 10 years, 20+ network service providers have established network presence and progressively built pan African network coverage for Enterprise Network Connectivity. The network footprint, portfolio and technicalities of the service offering however varies considerably with respect to number of countries and so does the pricing and quality of service. Hence a competitive commercial bid for pan African enterprise connectivity more often than not tends to lead to negotiations with more than one network service provider. This article attempts to demystify the landscape of Enterprise Network Connectivity across Africa and develop a clear picture on what global service providers and enterprise customers, African and/or Global, can and should expect.
Geographically the sub-Saharan countries in the East and West of Africa are grouped together are these countries are mostly served by the same set of subsea cables. However East and West of Africa has unique asymmetry, primarily because of number of countries. As a result several regional ring-protected fiber networks spanning 8-10 countries in East Africa provide a platter of connectivity options. This is not seen in West Africa, other than small network clusters spanning 3-4 countries. However, the affinity to connect to Brazil is noticiable. An East-West network corridor connecting Lagos and Nairobi or more practically Nairobi-Kigali-Muanda or Lusaka-Lubumbashi-Muanda remains a wishful requirement. The North African countries are generally not included in discussions about Africa as those countries have greater allegiance to Europe and Middle East due to geographic proximity and also served by a different set of subsea cables. The Francophone countries in North-West Africa are grouped together as difficult countries, due to sparse network infrastructure, few subsea cables leading to disproportionately high price for Enterprise Network Connectivity. Possibly the overhang of French imperialism that in due course led Orange to acquire significant shareholding in the incumbent providers in Francophone countries and thereafter its influence on the course of regulatory reforms, could have had a role to play.
The adjoining tabulation attempts to profile the network footprint of 18 network service providers across 39 countries. It’s based on information available in public domain and my experiential learning while driving partnership initiatives. The tabulation is evolving representation as network service providers continue to expand network footprint and populate new countries with PoP setup or extend coverage through NNI. The global leaders like BT, AT&T and Verizon are not included in the tabulation as not much is available in public domain about their Africa footprint, though it is not difficult to predict the likely countries looking at the tabulation. In this context a ground reality, that became evident in last one-year is, the pace of PoP setup across Africa has slowed down considerably other than lifecycle augmentation and upgrades. The focus now is to sweat the network assets. Further spreading the network built too thin with heavy spent does not seem to withstand increasingly stringent ROI scrutiny. Some network service providers are even contemplating winding up some of the underperforming PoP setups and trim underutilized backbone connectivity as cost saving measures which anyway is a global phenomenon. In the next 2-years, the new frontier of Enterprise WAN edge – SDWAN, will expectedy gain momentum as transport agnostic network overlay. The community of African enterprise customers led by African executives and entrepreneurs could be more open to MPLS replacement, much to the delight of the SDWAN technology vendors, as many of them may not have had the experience of and/or binding dependency on MPLS VPN. In this context Aryaka has taken the lead to setup PoP in South Africa.
The global network service providers and pan-African network service providers have pursued a three pronged approach to establish network presence in countries base on demand forecast and product roadmap. That’s of course no different from what they do in other geographies globally. In countries where they are licensed to provide service, it’s comparatively easier to build fully owned PoP setup. Thus providers that are primarily mobile network operators like Airtel, MTN, Vodafone/Vodacom, Orange and even Etisalat have had an advantageous position in bunch of countries. Added advantage is fiber and wireless infrastructure to provide Local Access connectivity till customer premise and in-country field operations staff. Another group of providers that are not mobile network operators also merit entry in this club. They are SEACOM and WIOCC as they have licensed entities in countries where the subsea cables land in Eastern Africa. Further they have developed in-land terrestrial connectivity to land-locked countries with Transmission, IP and/or MPLS PoP setups. It also includes Liquid, Simbanet that have developed fiber developed backbone and metro fiber network infrastructure from Kenya to South Africa and poised to move south-west Africa. This group also includes Paratus Telecom that has fiber network spanning 7 countries. Then there are providers that are neither mobile network operators, nor have invested in backbone or metro fiber infrastructure but have licensed entity operational in certain countries. These include CMC Networks, Internet Solutions and PCCW Global with PoP based coverage of 10-15 countries. In fact CMC Networks claims PoP based coverage of 53 countries, however that is usually taken with a pinch of salt by industry insiders. With smaller footprint of 4-6 countries it also includes Gilat Telecom (subsidiary of Gilat Satcom), Isocel Telecom, Alink Telecom and BringCom. The last in this category are country-specific providers that also provide long line cross-border connectivity to the neighboring countries. This is more prevent in Kenya, Tanzania, Uganda and South Africa. Given these providers have licensed entity in 4-15 countries, for other countries of interest they adopt a partnership centric approach for PoP setup.
Pursuant to partnership centric approach, the global network service provider identifies, evaluates and on-boards an in-country carrier partner. This is followed by contract negotiations culminating with contract signoff. It then gives way to network rollout and operational readiness. This end-to-end partnership PoP readiness could take anywhere from 8-18 months, depending on depth and rigor of technical, commercial and contractual engagement. In countries still shackled with incumbent monopoly, contract negotiations could be highly challenging exercise. Some of the remarkably challenging countries are Morocco, Tunisia, Senegal, Cameroon and Namibia. Thereafter the network presence is established as partnership PoP, also called VPOP. In VPOP setup the oversea network service provider shoulders the capex and opex required for VPOP setup as investment for network expansion. The role of the in-country carrier partner is to lend its telecom license for the PoP setup, such that the in-country carrier partner owns and operates the partner PoP and is the provider of service to customer entity in that country. However, given the partner PoP is part of the global network of the global network service provider, the PoP setup is centrally managed from the NOC. In this approach, the capex and opex necessary is invested by the network service provider interested to establish presence under the ambit of license of the in-country carrier partner.
The third approach is to establish NNI (network-to-network-interconnect) with multiple in-country or regional network providers with varying pan African footprint in London or Johannesburg. This is evidently the simplest and least cost approach. However this approach has certain disadvantages. The regional or in-country carrier partner’s network is not accessible or visible to the global carrier partner. Even though in case of managed network service the CPE at the customer premise is managed by the oversea network service provider, the carrier partner’s network remains a black box. Demanding enterprise customers therefore specifically ask for PoP based solution and insist that countries served through NNI arrangement with carrier partner is clearly identified. Pan-African providers are progressively more transparent to reveal the matrix of carrier partners they work with.
More often than not a global provider adopts a balanced mix of the three approaches to attain the desired network footprint and country coverage. This community of network service providers also collaborate amongst themselves with NNI setups to stitch up globally marketable pan-Africa network coverage. In simple terms they bundle or resell each other’s service to address pan African opportunities. Due to this the country specific pricing varies considerably, as not all the providers available have equally strong network presence in a given country. Hence global network service providers while developing commercial bid for pan-African opportunity attempt to get price quotes from 2-3 providers to negotiate the best price, unless the bundled deal is negotiated to be competitive enough as compared to peace meal solutions stitched together.
PoP setup is not end of the story for setting up network presence. We need to look beneath the hood and ask the right questions to evaluate the PoP setup. Usually with PoP setup we expect carrier class network equipment with fully duplicated cards for fail safe configuration. Hence high-end CPE grade router used for PoP setup does not serve the purpose. Further PoP equipment mostly being Cisco, Juniper or Nokia and possibly Huawei, it’s necessary to ascertain how is lifecycle service support managed to deal with equipment breakdown. It should be checked if things like Cisco Smartnet support is available in that specific city address or if the equipment vendors stocks spares locally with 24/7 4-hour on-site delivery. Further the PoP setup should have resilient backbone connectivity on diverse subsea cable systems or over terrestrial routes.
In addition to PoP setup, Local Access is needed from the PoP till the customer premise. The in-country carrier partner usually has the metro network infrastructure to deliver Local Access. Fiber connectivity with ring topology is usually available in the business districts. However well-documented list of connected buildings that could help automate in-house local access feasibility tools is difficult to get. Further it is relevant to enquire if the fiber is underground or areal. Arial fiber laid in huff and hurry to meet deadline could lead to worst of nightmares with service assurance. An alternate to fiber connectivity is wireless connectivity. These are point to point RF links over licensed or unlicensed spectrum that can be configured to scale from 2M to 20M. In some cases 3G/4G links are also deployed as backup links. The providers generally claim wireless connectivity tends to be more reliable than fiber connectivity as fiber is susceptible to cuts due to developmental or even destructive activities. Commercially too wireless connectivity is less expensive and hence preferred over fiber connectivity in some cases, where the customer is focused on the pricing and SLA, and not solution design and physical media. However wireless connectivity should be evaluated sufficiently to ensure technicalities of advanced WAN solution design are supported. Ideally, budget permitting, fiber connectivity backed by wireless links would lead to committed monthly uptime of 99.99%. A caution needs to be observed while dealing with Hybrid WAN solution with MPLS and Internet. Many of the network service providers use the same router to deliver MPLS and Internet. In that case, in addition to security concern, the PoP router becomes a single point of failure.
A closer look at the consolidated picture of pan African network footprint of different providers, reveals interesting insight on the in-country market landscape and extent of telecom liberalization. Maximum number of PoP setups are in countries that are in mature stage of telecom liberalization and have a regulatory framework that welcomes investment in network infrastructure built and create a competitive multi-provider play such that customers have multiple options to compare and choose. This is best seen in South Africa, Kenya and Nigeria that have 15+ network service providers that have established PoP. On the other hand that are countries that barely have one network service provider with PoP setup. Because of this 2M Local Access price has stupendous variance from $200-$400 MRC in 6-8 countries to as high as $3500 MRC in countries like Djibouti, Niger or Chad. This also means higher the number of providers with metro and wireless network infrastructure, higher the number of subsea cable serving a country, lower is the TCO for PoP setup, more is the number of network service provider going for PoP setup and eventually more competitive pricing for Enterprise Network Connectivity.
The PoP setup merits some more clarity. While MPLS PoP can deliver L3 and L2 MPLS service for enterprise customers, IP PoP setup is primarily intended to serve wholesale IP requirements of ISPs. Further Carrier Ethernet service based on technologies like PBB require separate PoP setup. However, demand of Carrier Ethernet service only felt in South Africa so far. Hence few of the network service providers have MPLS, Ethernet and IP PoP setup while most others have MPLS and/or IP PoP setup. For most of the African countries, the business districts are in the capital city. Hence the PoP setup is usually in the capital. This scenario is changing for some of the countries like South Africa and Nigeria that have three (3) industrialized cities and attract network service providers that are licensed in Nigeria and South Africa and target in-country enterprise customers to develop multiple PoPs. Some providers have 4-8 PoPs in South Africa and 3-8 PoPs in Nigeria
From this review, it is evident that pan African Enterprise Connectivity is quite well established with 20+ network service providers leveraging 20+ subsea cables to cater to 54 countries with 200+ PoP setups providing extensive network coverage, albeit with considerable variance of network footprint. There are 17 countries that have 3 or more network service providers and are coincidentally served by 3 or more subsea cables. On the other hand, 23 countries, mostly characterized by incumbent monopoly, that have 2 or just 1 network service provider that are either land locked or served by most by 1 or 2 subsea cables in some cases. Incumbent monopoly is irrefutably detrimental to telecom liberalization and flourishing of competitive landscape of telecom services to enable Enterprise Network Connectivity.
Nothing described in this article is intended to create a minefield of mixed interpretations. Constructive feedback from industry colleagues and proponents of Enterprise Network Connectivity are welcome to rectify any inadvertent factual inaccuracy that might have sneaked in.
(The views expressed by the author in this article are his own and not necessarily relates to the organization the author is attached to)