The 5G network transition is now underway to what many true believers would consider “real 5G:” the implementation of Standalone 5G with the cloud-native mobile packet core. While the initial rollout of 5G was focused on introducing 5G spectrum to boost capacity, 5G Standalone can unlock features such as ultra-low latency communications and advanced network-slicing functions. It comes at a critical time for the industry — when communication service providers (CSPs), in the teeth of an inflationary environment, are looking to monetize their 5G CAPEX outlays, while also realizing higher levels of service agility and operational efficiency.
A critical building block in the introduction of 5G Standalone is the massive technology shift associated with the introduction of the cloud-native 5G Core. Cloud-native or “born in the cloud” workloads are agnostic to underlying infrastructure, can scale elastically to meet demand and promise features such as in-service software upgrades, i.e., allowing for new feature introduction without network downtime. Cloud-native concepts underlie many of the most scalable software applications on the planet, from video streamers to social media applications. The expectations are high that they can usher in a similar revolution in telco cloud.
Unlocking the service agility and operational efficiency of the cloud-native paradigm, however, requires CSPs to transform into Cloud operators, an unfamiliar role. To understand the journey ahead, it’s instructive to look at a bit of recent history in the evolution of telco cloud.
The state of telco cloud and cloud platform options
The previous cloudification cycle in telco networks ushered in the era of Virtual Network Functions (VNFs). The VNF era focused on disaggregating network functions from their underlying physical infrastructure, but largely failed to deliver the anticipated operational and economic benefits. Reasons abound: the initial transition from physical appliances to virtual functions resulted in inefficient implementations. Automation, a key goal of cloudification, was dominated by vendor-centric solutions, resulting in complex integrations and siloed operational processes.
Shifting from VNFs to the Cloud-native Network Function (CNF) paradigm, then, is a “redo” for the telco industry, and a chance to get it right this time. The fact that this technology shift also underpins the transition to 5G Standalone networks only raises the stakes.
A key decision that CSPs must make as they roll out their CNFs is the choice of the underlying cloud platform and infrastructure. The telco virtualization journey up to this point has been marked by two broad alternatives for cloud infrastructure:
- A “DIY” cloud software solution from a vendor such as RedHat or VMWare, coupled with hardware, networking, etc. is often procured by the CSPs. In this scenario, the CSP assembles, integrates and manages the cloud infrastructure, pursuing a best-of-breed vendor strategy. However, it also requires a substantial investment in operational and cloud expertise.
- A second option is an integrated cloud infrastructure platform from network infrastructure vendors such as Ericsson, providing consistent and verified deployments. This can provide more rapid time to market and verification cycles as the network vendor often pre-certifies on the cloud platform. However, CSPs can lose some of the horizontal benefits of a single-vendor approach.
Hyperscale Cloud Providers (HCPs) such as AWS, Microsoft Azure and Google Cloud are relatively new entrants, bringing their substantial enterprise/IT cloud credentials into the telco cloud landscape. HCP cloud options are a different beast, providing not only an integrated cloud infrastructure platform but also bringing 100s of software services spanning building block capabilities such as automation and data management that provide the basis for a true DevOps platform.
HCP cloud has been transformational in enabling the rapid scaling of several software platform companies, e.g., Netflix’s use of AWS is a great case study. Advanced HCP tooling, the ability to rapidly scale infrastructure needs on demand, and DevOps principles have enabled these companies to introduce a torrid pace of features, and scale globally at phenomenal rates.
Can telcos leverage HCPs in the same way?
HCPs in the telco cloud: Early stages of maturity
While HCPs bring unquestionable cloud platform and operations credentials to the table, telco cloud infrastructure must meet many specific requirements: regulatory requirements such as data residency and lawful intercept, stringent 5 9’s availability requirements as well as high throughput and advanced network features required by traffic bearing network functions.
Broadly, these requirements are relatively new to HCPs, who are used to offering Service Level Objectives, rather than stringent, contractually bound Service Level Agreements. HCP Automation solutions that serve enterprise or IT functions well are not always good fits for complex Telco network functions.
The telco vertical, however, is a frontier in the competitive arms-race for cloud domination by the HCPs, and they are investing rapidly on telco-friendly features. A case in point: Telco cloud often requires on-premise cloud platforms deployed close to the Radio Access Network (RAN), given the latency sensitive nature of certain traffic-bearing network functions. To meet this demand, HCPs have rolled out a variety of on-premise or edge cloud platforms that can support telco requirements such as Google’s Distributed Cloud Edge (GDCE), AWS Outpost or Microsoft’s Azure Operator Nexus (AON).
CSP adoption: Many concerns underlie a cautious approach
CSPs have largely taken a cautious approach to the embrace of HCP clouds for telco workloads such as the cloud-native 5G Core. While HCP cloud platforms offer many attractive features, CSPs are carefully analyzing the TCO associated with HCP stacks. In some markets such as Europe, the embrace of HCP stacks needs to account for geopolitical and regulatory concerns. CSPs are also deeply concerned about the potential “lock-in” of the HCP cloud approach. Using advanced HCP cloud services can dramatically improve agility and feature velocity at the loss of portability. Switching from one HCP cloud provider is non-trivial and involves significant expense.
CSPs are making choices across the spectrum. Early adopters, such as DISH (which is building out a greenfield network), have chosen HCP cloud. Other CSPs are staying with DIY options (e.g., Verizon), or relying on a mix of network vendor provided platforms.
What happens next? Our crystal ball…
The entry of HCP cloud adds to the bewildering array of choices that CSPs are faced with on cloud infrastructure. CSPs are looking for consolidation and further standardization and portability in telco cloud infrastructure.
In this period of transition, CSPs need to carefully weigh overall TCO associated with the various cloud infrastructure options and balance it against their own specific considerations. Do they have the operational skillset to run their own cloud infrastructure? Do they have the facilities and transport footprint needed to optimize HCP cloud infra spend?
Looking ahead, there are some clear benefits to the adoption of hyperscale cloud for telco operators. The adoption of HCP cloud will likely initially be focused on use cases where the operational & commercial benefits can be easily proven.: Our customers have shown in interest in cap and grow use cases for 5G Core expansions, disaster recovery and redundancy configurations, and selective deployment of parts of the 5G Core network serving functions such as user management, exposure and policy.
Given the installed base of private & DIY clouds though, and current low volume of commercial HCP adoption for telco workloads, we believe the industry will be in a hybrid cloud situation for the foreseeable future.