By Kevin L. Jackson
The telecom industry is rapidly moving towards a future in which networks must have the capabilities of delivering services with the agility, flexibility and scalability only made possible by full programmability. More than software defined networks (SDN), these capabilities will deliver software defined service domains that can interact seamlessly with other independent service domains in a multi-tenant environment. The result will be a telecommunications industry capable of evolving and adapting to the emerging requirements of our networked society. This is why I was so excited to interview Henrik Basilier, Network Expert Network Architecture Evolution at Ericsson.
Since Henrik’s role with Ericsson is understanding and driving Ericsson’s strategy around the emergence of 5G and the Telco market horizon, I focused on asking questions around critical capabilities like network slicing and the network architectures this technology can enable.
Kevin: Thank you for taking the time for this interview Henrik. Before we get into the technology, please give me some insight into your role there.
Henrik: Thank you for having me Kevin. My role here at Ericsson Digital Services is a bit of an alternative career path. This is a leadership position with no direct reports where my role is to work with technology and architecture and look at the long term strategy for the company. In short, I make sure that we are using the right products in the right way from a strategic perspective.
Kevin: Clearly an important position at Ericsson. One thing I’ve noticed is that Ericsson’s strategy seems to have network slicing as a core focus. Why is network slicing important and in what context do you place it?
Henrik: 5G is all about opening up new business opportunities, providing services to new customers (e.g. industries) and devices. However, these new use cases and opportunities drive very differentiated requirements on the network and its operations. This calls for a new approach of building and operating networks, where the network is viewed as a set of logically isolated networks on top of a shared infrastructure, which can be separately customized, optimized and operated. This is what is referred to as network slices.
Kevin: What do these technologies do to the current networks?
Henrik: the network gets transformed from a monolithic “one size fits all” towards a horizontalized approach, where a shared infrastructure can host a multitude of separately managed network slices
Kevin: What specific network slicing drivers do you see in the marketplace?
Henrik: One main driver for slicing is tailored customer offerings. This is usually appropriate when companies are trying to address new marketplace needs and motivations. Another is the need for flexibility and agility which drives both topline and bottom-line improvements. Network slicing is expected to improve time-to-market and reduce total cost of ownership. And on a final note, network slicing allows for a reduced risk.
Kevin: Is network slicing primarily a new technology?
Henrik: Network slicing is first and foremost a transformation journey and effort towards the target of logically separated network slices. This is made possible by a set of enablers, some of which have been available and used for a long time. One can also claim that network slicing is building upon the idea of virtualization, but taking it one step further, to the complete network.
Kevin: How does this affect the companies themselves?
Henrik: This new approach requires an organizational evolution for many service providers, as it impacts how the CMO, COO and CTO will work with each other. In essence, it enables a more agile approach to service orchestration, something that the organization needs to take a 360-view on how to address.
Kevin: That said, how can we best get to a bird´s eye overview on the required future architecture?
Henrik: In the bottom we have a shared infrastructure. On top of this, network slices are assigned using set core network applications and partitioning of the Radio Access Network and the Transport Network. Across the slices, one management and orchestration layer delivers life cycle management, assurance and analytics, subscriber provisioning and service design. This layer is directly coupled by a Digital BSS layer, with order management, catalog management, Charging/billing and SLA tracking capabilities.
Kevin: Getting closer to what is required then, what are the essentials of the RAN and Core support for slicing?
Henrik: On the RAN side you need the capability of running core network aware slicing, with active selection of core network nodes and include both 4G and 5G slice identifiers. Second is the sheer Radio resource partitioning, allowing for a dynamic radio resource partitioning and service separation. The third essential capability is the quality of service guarantee, securing service separation within a slice. And last we have the capability of slice aware observability, using flexible counters for monitoring the usage load and performance of each individual slice. Across these RAN capabilities it is important to highlight that they are available today in LTE, using PLMN and SPID and interworking with the right designed core network. The same comment really go for the slicing capabilities in the core network. For these, the EPC standard encompasses a set of key characteristics. CUPS with a separate user plane per slice is one, the support of existing tools such as APN and DÉCOR/eDECOR are other. Then, as we move into the 5GC standard, the slicing capabilities are effectively further enhanced, through means of primarily DNN, S-NSSAI and PLMN Id. Through this evolution path, the core network plays its part in isolating and separately managing slices, optimizing edge-relying low-latency services, and sharing VNFs across different slices.
Kevin: Does automation lie on top of these layers?
Henrik: Yes. Automation is key to economically scaling network slicing. With hybrid networks increasingly becoming real-time, this is where service orchestration, assurance and analytics all require complete automation. Orchestration is really about configuring each slice for launch across the RAN, Core and Transport networks. We refer to it as model-driven orchestration, where full automation applies TOSCA templates for the fastest and most reliable handling. To that comes the subscriber and policy provisioning, that steers who, how and what to connect to any certain slice, meeting the conditions stipulated by the different SLAs. Service assurance and analytics capabilities, hook into all parts of the underlying network and continuously check on the state of the service and slice performance, through closed loop capacities. Service assurance and service experience management is really central. Visualizing, analyzing and automating the operations of network slicing is fundamental, and Ericsson has the technology and solutions to provide all of these capabilities.
Kevin: Thank you Henrik for sharing with us your insight on the future of networks.
Henrik: Great to work with you on this Kevin. Thanks.
For more on network slicing, view the Ericsson on demand webinar, “Network Slicing: Building essential end-to-end capabilities in 4G/5G” at http://bit.ly/2WgxAaf
Today’s CIO navigates the twin challenges of enabling new business models and managing rapid technological change. Cl...
Digital transformation infuses digital technology into all areas of an organization’s business or mission. Its fundam...
Q: Is it time for my company to jump on the composable architecture bandwagon? A: Composable architectures are quickl...
Workplace productivity is hurt every day by the very technology developed and purchased to improve it. Forrester ann...
Recently, Ericsson Digital released an amazing report on Edge Computing and 5G. In it, they explained how distributed...
An Interview with Henrik Basilier By Kevin L. Jackson The telecom industry is rapidly moving towards a future i...
Helping clients address the trends and challenges presented by the Financial Services industry is the main focus for ...
Next week, I will be influencing #LikeABosch as I accept an invitation from the company to attend Bosch ConnectedWorl...
First cloud computing then multi-cloud. How can we get ahead of this digital transformation nightmare? These are the ...