Everyone who needed a SIM is more or less having it, and in order to grow further, OpCos need to transform themselves from being just a Communication Service Provider (CSP) to a Platform Provider, offering any application to run on one of their network slices, accessible from any device, and from any location – fixed or mobility.

Cloud Service Provider is the new CSP
 

One of the first Operator to ready their telco cloud had to wait VNFs to fully mature. But this in fact provided them with an opportunity to start a new business line, selling cloud services to those businesses who preferred locally hosted DC. There are endless opportunities from hosting e-commerce, hospitality or campuses, to larger enterprises. Imagine an enterprise customer getting a single platform usage bill for fixed line, mobile, broadband, VPN, as well as for hosting their web/customer portal, or IT applications.

SaaS Opportunities for Telco SPs
 

An Operator’s existing voice and data services could be their first client, and even a competing VoIP provider can use the phone slice and vIMS, eventually sharing revenue with them, which otherwise would have been completely lost to IMs. 

PaaS Opportunities for Telco SPs
 

A content provider who intends to stream to tens of thousands of spectators in a stadium, can use the Operator’s platform to spin their streaming server at the edge of network, thus saving terrestrial network for the Operator, and hardware installation for them. And it won’t be long when web hosting would move to a distributed telco network from centralized data centers, particularly those requiring low latency.

Opportunities for Telco Vendors in the Cross domain competition

One of the first opportunities for telco vendors will be in networking, as the requirement moves from traditional hard to Software Defined Networking (SDN).

Eventually, telco vendor’s sale in future would rely on developing cloud native decomposed components, offering flexibility in purchase and shielding the threats from smaller IT companies, who are already writing VNFDs for basic functions. Even if lighter functions such as firewall or router VNFDs can be developed by small IT vendors, for complex network functions such as PGW, it would still be the telco vendors with their experience and regression testing, that would lead. Decomposing the VNF into VNFC would help Operators realize more business opportunities and ease of purchase, as and when required.

OpenStack is moving north, and not only NFVI, but with Tacker the VNFM and NFVO can be generic and open-sourced. While it would be a potential loss of business for telco vendors, it also creates new opportunities for them to re-design their OSS/BSS solution and offer new products that can be integrated with Open NFVO, for end-user controlled service automation and management.

And not to forget the huge opportunities of telco vendors in consulting and managed services, where they would have the upper hand compared to IT consulting vendors.

Connecting 5 dots of 5G

5G is not all about speed, rather it's latency that will shake-up the current hardware-locked architecture. Ultra low latency for mission-critical IoTs means the traffic needs to be handled at the edge of network by functions currently residing back in the Core. 

NFV
Unlike current hardware-locked Gateways and PCC nodes, Virtual Network Functions (VNF) needs to be spawned from component library, and run on stateless process units in any corner of the network. For example TDF function at the edge of network identifies corporate, TV or electric grid traffic and breaks them locally instead of bringing them all the way to the Core and routing them back to their respective servers which might be located at front. In fact most low latency, mission-critical traffics such as V2X needs to be routed back from the front to another vehicle, human or traffic info database with their server functions virtualized and distributed at the edge of the network.

MEC 
Delivering higher bits on air using unlicensed band is not the challenge, but it's to bring them to the Core. Gigabits ANs placed at every 50 meters or so, would flood the Backhaul. Thus Mobile Edge Computing (MEC) is a must to have, breaking as much of traffic at the edge as possible for example UHD, V2X, corporate and even WLAN with ePDG anchoring for seamless voice over IMS. It's not feasible to carry the data flood all the way back to Core by scaling hundreds of miles of fiber by several folds.

SDN 
One of the key to success would be to fit in the right flavor of SDN and match it with Operator's specific requirements, given the fact that ETSI alone has recommended 13 flavors of SDN. The platform offerings would rely on network slicing and end-end service chaining using three SDN controllers – Infrastructure, WAN and Tenant, all governed by a single network-wide Orchestrator.

SFC
Existing networks have single GiLAN path with all types of traffic hitting all nodes, which unnecessarily loads the Network and Value Added Services. SFC controller chains various elements on the GiLAN, so that traffic can be steered to their respective nodes. For example streaming doesn’t need to go through vTCPO, or TCP traffic need not hit the vVideo Optimizer, thus bringing network efficiency.

CUPS
Some of the least spectral efficiency technologies such as GSM would eventually shut, but others such as LTE using the vacant 900 MHz spectrum would still be needed to provide long range  coverage for mobility and session management, while the small cells would provide the much needed thick pipe-on-air for user plane traffic at every 50 meters or so. Now this put another requirement for control and user plane splitting (CUPS), which obviously is linked with the Micro-service Architecture (MSA) in which the network nodes are split into smaller VNF components (VNFCs). 

Ramping-up the Radio
 

Do we need to wait for standards, devices, or radio readiness of VHF/UHF bands, before kick-starting the build-up towards 5G? Not really! 

Remember the last evolution, when Operators started evolving their Core to 4G EPC, even before they had the LTE spectrum. This time the Core, Backhaul and Fronthaul all needs to be prepared to handle the arrival of data flood over very wide unlicensed spectrum. Moreover, virtualization of Core is needed to bring in the much needed elasticity and agility and to automate service creation and management, regardless of the radio pumping 100Mbps or 1Gbps.

So why wait for VHF/UHF readiness, instead why not start with the existing unlicensed WLAN, using any of the aggregation technologies - LWA or MPTCP. Even without aggregation, 1Gbps can still be delivered as pipe-on-air from single AN, supporting LTE-A and WiFi 802.11ac. Users can connect over SIM/Non-SIM devices, doesn't matter, as long as the capacity on air is 1Gbps/node/50m. Moreover we don't expect user applications exceeding 50Mbps anytime sooner, but the idea is to create the aggregated capacity on air for high speed data, UHD, and IoT.

Service Oriented and Multidirectional Traffic Architecture
 

And finally, it’s time to redesign the traditional west-east traffic flow architecture, to a multi-directional one, with the edge of network flows directed towards Core in both west-east and east-west directions. Additionally breaking-out north or south, thanks to MEC and MSA decomposing the current hardware-locked nodes into VNFCs that can be spawned on demand in any corner of the network.