The path to a globally available hybrid cloud for 5G Core

While there are many reasons for a CSP to consider using HCP infrastructure for its telco applications, there is an equal, if not higher, number of factors to consider when it comes to the feasibility of such a decision. In discussions with CSPs, a number of areas are consistently being raised. While these are shared in broad and general terms with businesses in other industries, there are also telecom specific perspectives that are  worth mentioning.

 

Use cases and business priorities

There is no shortage of examples where a startup rapidly and successfully transformed an idea into a multi-million business by utilizing HCP public cloud platforms. In many of these examples, the startup was not only able to develop and scale its service with the uptake of a growing customer base, it was also able to extend its global reach to other continents. Often, these startups describe their industry purpose as creating new end user services, as opposed to developing technology platforms and investing in private data centers.

However, there is also evidence from other ‘internet based’ success companies who attribute their fortune to strictly keeping its technology and knowhow within the company, choosing to evolve its own internal infrastructure capabilities and footprint as far as possible. According to this approach, it is only  by utilizing and growing their own technology capabilities that they are able to do what they do. They stay relevant by gaining insights into how their service can be improved by analyzing information within their own private platforms. This becomes especially important if they are concerned about future competition from HCPs’ own developed services.

In several cases, the cost for running services solely in a public cloud infrastructure has reached unsustainably high levels, negatively impacting the market value. As a result, alternative platforms need to be found.

For many companies, the way forward is a hybrid cloud where the combined strengths of private and public platform capabilities make the most sense.

But how should a hybrid cloud be used? Which types of applications should run in a private data center, and which are better placed in an HCP’s public infrastructure?

From a global perspective the circumstances impacting a CSP’s hybrid cloud choices varies.

• For a greenfield CSP with a need to rapidly install and grow a Mobile Core network, it could be tempting to place as much of the applications as possible in an HCP platform. This would be especially interesting if cashflow is constrained or limited during the buildout and a ‘pay for usage’ business model is applied.
  However, the needs of the CSP might change over time. Business priorities might delay the solving of    any deficiencies and scalability issues, where the need for utilizing private infrastructures could prove necessary.
  Important to remember is that two different greenfield CSPs could choose different strategies depending on various factors, such as the availability of HCP footprint in the respective country, regulations and which competition they are facing.
  
  
• An existing CSP with an already established Mobile Core network could look at the complementary opportunities from the start. In this scenario, the benefits of a hybrid cloud are more related to the development of new and evolved use cases, potential revenue improvements and cost efficient deployment scenarios.  

Directional strategies are common in many industries where a C-level decision is made to either move the business in, or away from, a specific direction.                              .      

These directional strategies tend to have decade/semi-decade long targets, e.g. by the end of this decade target percentage of our business applications will be on a cloud. This longer time horizon ensures that challenges are not only related to technology and development, but also include operational and industry best practices under the influence of partners and regulators. Interestingly, success is often founded in a solid understanding of both its own, as well as its partners’ capabilities in relation to local market forces, paired with an ability to adjust along the way.

Secure telecom performance

The telecom network as we know it today is the result of years of evolution from a technology and network topology perspective. Every generation of mobile services was designed with an ambition for change, as well as being founded in existing networks. As is often the case, there is a drive towards the potential value, while also recognizing the cost of introduction. After all, if the cost of change is too high, the value might not be feasible for introduction.

Telecom performance or ‘telco grade’ is one of the cornerstones of the telecom industry and, over the past three decades, it has transformed from a fundamental industry requirement, to being shunned in the era of online services and OTT players, only to regain its reputation as a vital necessity during times where many services have become ‘society critical’. Online banking and online health care are two prime examples that clearly illustrate the importance of a stable mobile broadband service that is there when we need it and has the required security mechanisms implemented.

With the broadened use of HCP platforms and online enterprise conference services, there is an increased awareness of the consequences when underlying infrastructure and applications do not perform as expected. In some cases a loss of sound or video is experienced while in other cases, the entire service might be down for hours.

Telecom grade performance is a matter of expertise, experience, intuition, imagination and respect for human limitations. OTT network problems experienced today are similar to the problems experienced and solved by the industry years ago, despite the technology being vastly different. However, there is now a much better recognition of the need for telco grade and a willingness to understand how the necessary solutions can be developed in collaboration with relevant stakeholders.

Leveraging existing cloud investments and optimizing TCO

The telecom industry is especially known for its long investment cycles. Services often need to be made available on a nationwide scale with a focus on ‘coverage’ as a first step and ‘capacity’ as a second step. With a distributed network architecture it is often required to dimension the individual network sites in one, three and five year perspectives. The first installation should be able to manage the capacity growth until year three which, after expansion, should then manage the capacity needs until year five and so on. Due to such long term investment cycles, telecom is, by tradition, a CAPEX oriented industry where assets are depreciated over several years. This means that any infrastructure in the network not yet depreciated is not really subject for replacement.

There is an expectation that, when HCP private and HCP public infrastructures (our first blog post, Collaboration is key,  discusses this in more detail ) are added to an existing infrastructure to form a hybrid cloud, it will be possible to lower the total cost of ownership (TCO). While any TCO-reduction would depend on, amongst others, the business models and price levels of the hybrid cloud resources and tools, there is also a need to consider changes to the network topology, link capacity, power and cooling principles to mention a few.

The hybrid cloud TCO is dependent on different factors which, combined, could lead to very different results depending on where  a CSP is geographically located. Examples include:

• The topology and size of the CSP network
• The size of the country
• Location of HCP cloud regions and distributed clouds
• Country -specific HCP price levels and service offerings

Open and sustainable cloud strategy without lock-in

When discussing lock-in, it is important to note the various meanings of this term.

Risk of not being able to move to another cloud provider.

In their risk assessments, many of the larger, well-known ‘online’ media-, streaming-, travel- and delivery service providers who today run most of their applications on HCP clouds, state that they have become dependent on a certain cloud provider. Moreover, it is not easy (if at all possible) to change to another HCP due to the way their software interacts with HCP specific and proprietary cloud services and APIs. The business risks also include uncertainties regarding future price levels and what might happen if the cloud provider would prioritize its own competing services.

Skills and people

There are concerns regarding the irreversibility of moving from a private cloud infrastructure to an HCP infrastructure in terms of skills and people. Many businesses who have made similar moves in the past, only to realize that they need to revert back to an inhouse private cloud, soon realized that the people with the right skills are no longer easy to come by. Although the HCPs have certification programs for people to develop their skills, these candidates remain scarce and are not easy to come by. A CTO in the financial industry likened the dependence of finding HCP certified/skilled people outside the company as ‘joining the war for talent out there’. This is in contrast to controlling its own technology and intellectual property, where it would be possible to fulfill an internal target where 80 percent of open positions could be filled by developing the skills of internal applicants.

Service contingency

There is a concern that society critical services become too dependent on certain sub-suppliers. Regulators and oversight organizations are working on clarifying their need for service contingency plans with  several perspectives needing to be addressed. One such issue relates to how contracts and SLAs are worked out between different technology players. This becomes especially important when considering that a technology relation can be formed by simply signing up for a service through an API. In other words, there might not be a formal contract in the traditional sense that, for example,  stipulates what should happen if the sub-supplier cannot fulfill its promises. Another issue is the risk of a sub-supplier becoming a ‘persona non-grata’ due to political reasons or misconduct. A converse scenario could also occur where a society critical service provider is not allowed to use the service of a specific sub-supplier.

Security and sovereignty

As stated before, telecom services have become society critical with normal life and important businesses not being able to  function in its absence.

While businesses are eager to explore the capabilities of a hybrid cloud, there are also concerns on a country level to preserve the security and sovereignty of provided society critical services. There are several perspectives including, but not limited to:          

• Who has access and can read my data?
• Will I have access to my own data at any time, and do I have control over where it is stored?
• What does the full list of partners and sub-suppliers look like, and can I trust everyone in the chain?
• Do I rely on infrastructure and services performed in another country?
• Can the service run and evolve based on access to domestic infrastructure, software and management tools only, or am I dependent on non-domestic services and partners?
• Which obligations do partners have to the jurisdictions they belong to?           

The answer to these questions varies across the globe and requires legal knowledge about both domestic, as well as international law, bilateral agreements and best practices.

It could be said that security and sovereignty is a moving target.

• While data protection through encryption could be a solution for now, future technology could potentially be used to break the encryption and decrypt the data at a later stage.
• Regulations are evolving and best practices defined and agreed today might not be fully compliant in the future.

Operational model and RASCI

One of the biggest concerns when considering a hybrid cloud is the potential impact on operational models and responsibilities. The way in which infrastructure, applications and services are collectively set up and operated today is well understood. The chain of command and response in case of network incidents is clearly specified between involved parties throughout the lifecycle of inc,luded technologies and tools.

A hybrid cloud, where HCP platforms are added to an existing network, creates the need to rely on another party to manage the infrastructure on which these society critical services run. The CSP is still expected to fulfill its obligations to customers and regulators, but depends on the HCP for the resilience of the service provided. Who does what, and under which circumstances are equally important to understand and agree. While a RASCI matrix is a start  to clearly define  who, at any given time, is Responsible, Accountable, Supporting, Consulted and Informed, a feasible operational setup is still required.

The path to a globally available hybrid cloud for 5G Core

The path to a hybrid cloud is a multiyear journey and we are only in the infancy of using HCP infrastructure and capabilities for telecom applications.

Many of the concerns described above are today addressed in isolation and there is a need to establish a way of working where these are collectively progressed towards a clear target.

Exploration via PoCs

We are currently in the exploration phase where a focus is placed on the technology aspects of telco grade characteristics, performance and security. This is being done together with CSPs and HCPs where a particular emphasis is placed on application and infrastructure requirements, as well as solution blueprints.

Operational and business understanding

We anticipate the next phase will include a deeper focus on the operational and business aspects with a particular emphasis on developing use cases and best practices to efficiently operate these. This is the phase where the hybrid cloud will have to prove itself from a TCO perspective with a clear understanding of how to best allocate and distribute 5G Core applications within the hybrid cloud. In other words, how to best utilize the combined capabilities of CSP private, HCP private and HCP public infrastructure.

Hybrid cloud for 5G Core certified and globally available

The third phase will address the Security and Sovereign aspects of a hybrid cloud for 5G Core. Regulations will evolve in different parts of the world. The EU will likely be one of the major drivers of both specific member state legislation, as well as EU level regulations considering the amount of member states addressed.

This is the phase where a combined  solution will start to emerge, one that is technically feasible and secure, proven from an operational and business perspective and fulfills both domestic and global legislation and regulations.

Ericsson is leading the way, engaging with CSPs and HCPs, aiming for a hybrid cloud for 5G Core which is Telco grade, Secure, TCO optimized and globally available.

 

Read more on the topic: 

Read up on my previous blog! 

More information on 5G Core