Posted on Reading Time: 7 minuteson
At the MEF 3.0 PoC Showcase in Q1 2021, the PoC (131) team was awarded “Market Game Changer” for demonstrating how CSPs can accelerate adoption of multi-edge cloud services through the orchestration of applications and connectivity on open edge compute platforms.
This blog post provides a selection of questions and answers resulting from the MEF 3.0 PoC (131) presentation. (They have been edited for clarity.) A recording of the entire showcase is available on demand. Our thanks to these team members for providing the answers: Wenyu Shen (NTT) and Susan White (Netcracker).
What is the definition of a multi-edge cloud service?
Multi-edge cloud services are services that need to be deployed on open edge compute platforms to meet specific demands, including latency, throughput, and data protection. Edge cloud platforms can be located on the enterprise premises or at the network edge and are managed by the Communications Service Provider (CSP). The services can include connectivity services, such as SD-WAN, security and private 5G, as well as applications, including high-value MEC applications such as IoT processing and enterprise self-developed IT apps.
In 2021, what is the biggest challenge for service providers deploying compute resources for an enterprise customer across public and private edge clouds?
At their core, edge compute platforms are meant to be derived from a suite of open components that are offered by a wide range of ecosystem providers. From a business perspective, this strategy makes sense for CSPs wanting to maximize their revenue opportunities with an array of use cases and a competitive vendor ecosystem focused on managing infrastructure costs.
However, from a technology perspective, integrating a large set of ecosystem partners can be challenging. And if you consider the myriad of different open source and standardization efforts that are driving the edge compute space, the path from service concept to service deployment can be long and potentially painful. Therefore, it’s easy to see the importance of driving proactive integration activities, like those offered by MEF, to allow the vendor community to collaborate and work through as many issues in advance of specific customer projects.
How does MEF LSO enable orchestration of a service across both public and private edge clouds?
MEF LSO provides a comprehensive and flexible framework for service lifecycle management that can easily be applied to public cloud platforms: placing the initial order, executing service configurations, and subsequent service state modifications. By further alignment with MEF LSO, CSPs and hyperscalers can accelerate the adoption of a multi-cloud strategy with standardized E2E orchestration.
Business-intent orchestration, merged with the MEF LSO framework, gives CSPs a comprehensive platform on which to build a cloud management solution suitable to manage both pure edge enterprise services and innovative MEC-based 5G slicing with multiple CNFs (Cloud Native Function). MEF LSO standards’ evolution towards vertical enterprise services would ensure the necessary orchestration platform features are supported to simplify the implementation of these service types.
Why are CSPs and hyperscalers working together in the multi-edge cloud service industry?
CSPs and Over-the-Top Providers (OTTs) should collaborate rather than compete with each other in the edge computing context. One of the challenges that the edge computing industry is facing today is the huge initial investment, which emphasizes the importance of partnerships between CSPs and between CSPs and OTTs. In this PoC, we successfully showed the possibility of several collaboration models. We deployed an OTT IoT solution on a CSP Edge, and deployed a CSP SD-WAN service on an OTT edge. We also adopted the latest network service to show the interconnection between the CSP and OTT networks. We also showed the possibility of orchestrating different edge solutions coming from both OTT and CSP domains.
In what way will CSPs evolve the multi-edge cloud service over time?
We have seen a paradigm shift towards hybrid IT, in which hybrid cloud and hybrid WAN have become mainstream. With hybrid IT, there is always a core edge sitting inside the CSP’s network, which seamlessly connects the WAN to the multi-cloud interconnect. A variety of value-added functions, such as SD-WAN gateway, VPN proxy, UTM, CDN, etc., can be deployed on it. In addition, enterprise users are also asking for upper layer functions, such as AI, IoT, gaming, self-driving, remote control, etc. To guarantee low latency for these applications and reduce WAN costs, they need to be located at the network edge and even the on-premises edge.
To keep up the pace of this transformation, edge computing, together with 5G, has become the main investment focus for many CSPs. Multi-edge cloud service will appear in the market soon and will be built on a distributed infrastructure across core, access, and on-premises locations. Enterprises will be able to select their own services, deploy to the right locations, and connect end-to-end with guaranteed networks.
What are the top three lessons learned during the development of this PoC?
It is possible to merge a very diverse set of technologies under a unified market-ready service in weeks. Each vendor in the PoC offered an impressive set of technical capabilities with a very high level of automation and integration potential.
An ecosystem-based approach for technical collaboration among vendors is essential to overcome integration issues, resulting in a market solution that minimizes configuration effort.
An evolution towards enterprise and edge cloud services requires an evolution of standards that take enterprise-specific business needs into consideration, along with ITIL-based practices that were developed to support standalone Enterprise IT systems operation.
Why is it important to provide branch connectivity as part of a multi-edge cloud service?
In a fully managed multi-edge cloud service, the end-to-end application experience is extremely important and must be assured at any time. In the delivery of remote services, many issues occur at the LAN and WAN connectivity, and that is the main reason CSPs should deliver and control the SD-WAN, security, LAN, and Wi-Fi infrastructure, as well as the private 5G.
The role of managed SD-WAN: Provide the best connectivity for each application across multiple WAN links. Monitor the different traffic flows, make changes to adjust SLAs, and identify the source of performance issues.
The role of managed LAN and Wi-Fi: APs and switches across a branch (factory, airport, etc.) provide critical connectivity to people and devices that require permanent access to edge-cloud services. The multitude of equipment connected, and nature of traffic, must be monitored to identify and report misconfigurations, potential strange behaviours, and malfunctioning devices or cables.
The overall branch management: Offering the LAN and WAN branch connectivity as a managed service represents a challenge on its own to the CSP. To manage and control the quality of the experience of thousands of connected devices, the use of AI and ML is extremely important to rapidly identify problems and self-resolve to scale the support operations.
What functions reside in WiFi access points on-premises? Are they considered uCPEs? Do they host VNFs or CNFs?
The WiFi access points are not considered to be uCPE. This is the role of the on-premises Edge Cloud platform on which the VNFs and CNFs reside. The APs provide radio access for IEEE 802.11ax (Wi-Fi 6), virtual Bluetooth LE, as well as a port to control IoT devices. The APs are connected to switches with single-click activation and auto-provisioning. The APs work in conjunction with AI to collect and analyze metadata in near real-time from all wireless clients, enabling rapid problem detection and root cause identification with predictive recommendations and proactive correction.
Which multi-edge cloud service use cases are already in production?
Today, we are seeing numerous customer use cases that enable deployment of edge compute infrastructure at the enterprise premises or network edge. These use cases include:
- uCPE: Provides software-based network functions on a commodity, open compute platform, hosted at the enterprise premises.
- Tenant space/managed edge: Provides enterprise workloads in either a customer managed, single-tenant or multi-tenant environment, hosted at either the enterprise premises or at the network edge.
- Wholesale: Allows managed service providers to enable one or more wholesale partners to resell edge compute capacity to enterprise tenants.
These use cases are being adopted across a number of industry verticals, such as smart cities, banking, healthcare, hospitality, retail, oil/gas/energy, and railroads, and typically include one or more of the following workloads:
Managed networking or communications services:
- Routing or SD-WAN
- Advanced security, firewall, UTM, or SASE
- WAN optimization
- Voice and data services
Enterprise value-add applications, such as:
- IoT sensor data collection and processing
- SCADA applications
- Video processing and analytics
- Enterprise self-developed applications
What is the minimum set of functions that need to reside in the on-premises edge in a multi-edge cloud service?
Most on-premises edge compute solutions require some form of managed networking services, like SD-WAN, routing, or wireless connectivity. Beyond that, other applications are unique to the specific customer use case. However, regardless of the workloads being managed, on-premises deployments require a high degree of automation and remote management capabilities, due to the distributed nature of those deployments. The edge compute platform must provide a common platform for running virtualized and containerized workloads, as well as a robust management paradigm supporting simplified deployment, service creation, day-N management, troubleshooting, and upgrade of the solution.
Which MEF standards were used for this PoC?
- MEF LSO Legato, LSO Interlude and LSO Presto interfaces were used in the PoC to order and provision enterprise services.
- A set of MEF W117 SASE guidelines was used to model secure overlay connectivity on top of the CSP’s network.
- The MEF W68 Cloud Services working draft was leveraged for public and private edge cloud service modelling.
- The MEF 70.1 draft standard was applied for the SD-WAN service to stitch enterprise and CSP network edge locations.
What was learned in this PoC that should be applied to enhancing LSO Presto and LSO Interlude APIs?
Enhancements to the SD-WAN service model (MEF 82), together with further definition of the SD-WAN service, would be very beneficial to decrease integration efforts by providing a blueprint.
For MEF LSO Interlude, it is critical to ensure that partnering parties leverage catalog-driven orchestration systems supporting a standardized set of service states and transitions. Multi-tenancy support is an important topic that should be developed further to provide enterprise-centric services. Service assurance scenario enhancements provide another field of further standards development.
What types of challenges do you foresee in providing enterprise edge service assurance when CSPs scale up and have issues with competing resources and priorities?
To satisfy the expected application experience, enterprises require strict SLAs on the end-to-end service and any issue should be located and fixed rapidly.
At the wireless, wired, and WAN domain level, AI-driven assurance is necessary to anticipate and react to any network issue. To satisfy the application-level SLAs, it is important to monitor the network quality end-to-end, from client to cloud, and this can be challenging for the CSPs to provide service assurance for edge services at scale.
Read more about: MEF 3.0 PoC (131) Orchestrated Multi-Edge Clouds
Other PoC (131) blog posts:
- Part 1: Building a Vertical Business Around Edge Cloud Services: Highlights from MEF 3.0 PoC Showcase 2021
- Part 2: Building a Vertical Business Around Edge Cloud Services: Highlights from MEF 3.0 PoC Showcase 2021
This PoC was presented at the MEF PoC Showcase in Q1 2021. Watch the Presentations on YouTube
About MEF’s PoC Program & Showcase
The MEF 3.0 Proof of Concept program effectively fosters innovation, seeds new MEF standards and projects, and accelerates our existing work within the ICT industry by providing our members—including service providers, technology suppliers, and other stakeholders within the ICT industry—the opportunity to collaborate on MEF 3.0-based use cases throughout the year.
Initiated by MEF members and facilitated by MEF staff, each MEF 3.0 PoC receives a unique, identifying number that remains unchanged as its title and messaging evolves over the life of the project.
PoC work is highlighted in public showcases and award presentations that explore individual Proofs of Concept. Learn more about these enabling technologies and the MEF 3.0 PoC Program.