This article is written about the Telecom Service Provider’s vision and roadmap for adopting the different types of cloud enablement and deployment strategy for cost efficiency and TCO ( Total cost of ownership) reduction, shorter time to market and making operations leaner. Cloud enablement use case examples include data center modernization, data center extension, application lifecycle management, and access to services.
Telecom service providers are both consumers and providers of cloud infrastructure and its services as they consume the cloud infrastructure services to run the IT or telecom applications on their network, and also extend cloud services to small-scale industries for running their applications. It is estimated in the next three years, telecom enterprises and their data center planners need to prepare for a major shift in the workloads that will drive data center expansion to support the increasing workload requirement.
There are four major data center cloud strategies which Telecom Service Providers are using or wish to adopt in the future. Same depicted in below diagram with different domains of the operator network.
a) Running Multiple (NFVI/MANO) from multiple vendors with bundled applications software. As cloud technology started to roll-on, this was the first option used by all vendors to run their application on respective clouds, so currently, it is the most deployed option in the operator’s network. Although OEMs have done customization to NFVI to get maximum performance benefits of hardware with new features of container management, SDN integration in NFVI, however applications and NFVI are tightly coupled to supplied hardware by OEM.
Managing multiple hardware supply chain has been a pain for SPs as they are bound to buy and deploy multiple cloud platforms under this option. In addition, there are other project execution issues which they have to face. For example, one vendor arrange hardware within a week and others could not do it for a month, it adds an unnecessary delay in overall project execution when it has hardware from both vendors. Needless to say, resource competence is another challenge to manage different types of hardware and NFVI platforms. Either they have to hire multiple vendor resources or hire an SI company to manage multiple cloud environments which is always costlier than other options being discussed later.
On NFVI market share, since NFVI and application are sold bundled, it’s market is directly proportional to the market share of the respective application. If Cisco has 30% market share of vEPC, so they have the almost same share of NFVI in the vEPC domain. This option is most profitable for incumbent OEM vendors ( Ericsson, Huawei, Nokia, and others) due to large profits from hardware, but due to operator inclination towards other deployment strategies, profits are getting a hit apparently.
b) Common NFVI and MANO from single OEM to run multiple application. This is the best strategy for operators who wants to reduce the cost, but don’t want to invest in their own development centers. They can choose one common NFVI platform from a single vendor based on features, price, standard compliance and run different network applications on it. This will help to reduce expenditure by giving bulk orders with higher discounts from one vendor’s NFVI hardware and software. It has other advantages too of shorter time to market, reduced maintenance, less supply chain issues, etc. This is easily achievable for a Service provider by arranging a lab of decided vendor NFVI and to test application from multiple vendors. Many operators are trying this option which gives benefits of cost reduction and shorter time to market.
HP and Dell are leaders for supplying hardware orders. MANO vendors like VMware-vCloud Director, Nokia-CBND, IBM-ALM are trying to get orders for common orchestrator ( MANO).
c) Inhouse development of NFVI and MANO using X86 and open-source software: the Third option which is being tried by few Tier-1 operators like AT&T(US), JIO ( India), Rakuten ( Japan) are developing their inhouse NFVI and MANO. This results in a range of approaches, from using ODMs for building one’s own systems to leveraging open source components for container and microservices. In either scenario, the SPs are procuring processors (often with specific customizations) directly from Intel or its competitors, storage capacity directly from HDD and SSD manufacturers, and next-generation network technologies (e.g., 25/50GbE, 100GbE, and 400GbE) from network component and ODM suppliers. AT&T is using Kubernetes and OpenStack infrastructure rather than VMware to provide greater flexibility and agility to support AT&T’s 5G network.
The O-RAN Alliance, founded by AT&T, China Mobile, DT, NTT Docomo, and Orange, recently gained new members, including Verizon, Ericsson, Reliance Jio and TIM. The O-RAN Alliance is focused on evolving radio access networks to make them more intelligent and open, including by leveraging real-time analytics, machine learning systems, and AI. Costs to deploy white-box hardware are significantly lower than deploying proprietary hardware, which will help operators realize savings not only on the hardware itself but also on the implementation and support of the hardware. Reduced complexity of installation and support is a key reason why the white box is less costly to deploy compared to proprietary systems. Rakuten estimates the cost to be around 50% lower when compared with a network using traditional architecture.
Off course there is no doubt on the requirement of competent RnD resources to run this option. But investing in this can bring revolutionary future results for operators. This is possible that Tier-1 operator will become a competition to existing OEMs after developing their own RnD centers. Acquisition of Radisys by JIO, Pjazr by JMA Wireless are recent steps towards in this direction.
d) Last but not least strategy is moving the application on public cloud and making Hybrid cloud for Telecom Infrastructure. Hybrid cloud solutions are designed to modernize on-premises data centers via access to public cloud services behind the firewall. Specific solutions aim to provide integrated access to private and public cloud environments while easing workload migration between both environments. This option is just taking feet of ground, however, the companies like AWS, VMware, Google are showing great interest to enter this arena with their matured and advanced cloud platforms.
Small OEMs are doing collaborations with public cloud providers to offer cheaper and feature enriched applications to SPs. Major advancements, in this case, are Optiva BSS launch on GCP ( Google cloud platform), MyCom OSI and Affirmed Networks vEPC launch with AWS. Major ice-breaker would be recent collaboration among ‘Vmware’ and ‘AWS’ which will give high-end private cloud services with most advanced containerized services access of AWS. AWS is very soon launching service OutPost too which is on-Prem AWS cloud. I think AWS could be a great challenge to incumbent OEM players and I will talk about its details in coming articles.
This option with a combination of option-c where SPs are trying to use x86 hardware will be most modernized and futuristic data center extension approach. It is highly cost-optimized, flexible, cloud-native and in open-source direction. Network functions traditionally run on purpose-built network appliances will increasingly shift to x86-based hardware, which will house and provision VNFs. Public Cloud providers being expert of scalable, elastic and cheap storage, ML/AI solutions will add-on to reduce cost and fast delivery. Apparently, this is going to be future of Telecom Infrastructure.
Summary: In my view Service Providers should move from option-a to any of b-d to bring TCO reduction, bringing uniformity of NFVI platforms, lean operations, faster time to market. The mix of option c and d should be adopted as soon as possible by them. Needless to say, incumbent OEM like Amdocs, Nokia, Huawei, Cisco, Ericsson who have more than 1B$/per quarter revenue from Telecom Infrastructure servicesˠ will get a hit by this change. The disaggregation of hardware and software threatens historical profit pools and forces incumbent vendors to invest in transitioning their businesses to support Everything as a Service (XaaS) and other business models, such as pay as you grow, which in many cases require scale and long duration before attaining profitability. They should carefully balance their shift from the old world to the new reality,whereby hardware upgrades are smaller in scale and targeted, and innovation migrates to the software layer. Companies such as Cisco, Juniper, Nokia, and Ericsson are becoming software-centric, though they are doing so gradually as this trend will take years to fully play out. Off course, SI companies like TechM, Wipro, Infosys, Capgemini will be benefited due to this change as scope of system integrators will increase year on year.
NFVI: Network functions virtualization infrastructure (NFVI) includes hardware (predominantly data center equipment such as servers and storage) and close-to-the-box software such as operating systems and infrastructure management platforms that compose the physical layer on which VNFs are housed.
VNF: In a Network Functions Virtualization architecture, a virtualized network function, or VNF, is responsible for handling specific network functions that run in one or more virtual machines on top of the hardware networking infrastructure, which can include routers, switches, servers, cloud computing systems and more.
MANO: Software platforms that manage and orchestrate network resources in a virtualized environment; management and orchestration (MANO) leverages automation and is responsible for policy management.
SDN: Software-defined networking (SDN) is an approach to networking in which control is decoupled from the physical infrastructure, allowing network administrators to support a network fabric across multivendor equipment.
ˠ Telecom Infrastructure Services include all Capex and Opex by service providers (wireless, wireline, cable, satellite and cloud operators) on or related to communications and IT infrastructure. four main segments: Deployment Services, Maintenance Services, Professional Services, and Managed Services.