Reminder

To have a better experience, please upgrade your IE browser.

upgrade
If you need help, please click here:

Etisalat — 2020 Landscape (Part 2 of 2)

Network Functions

As outlined in Part 1, Etisalat is striving to achieve a 60 percent virtualization of network functions by 2020. We summarized the market drivers for moving to virtualization, discussed the reasons behind the move towards Software-Defined Networking and Network Functions Virtualization deployment, and introduced the three key pillars that are the foundation of this future network architecture:

  • Software-driven, data center-based platform
  • Virtualized network functions and applications
  • Agile and orchestrated operations

We also detailed the primary goals for Etisalat’s future network: automated, scalable, and based on an open source platform.

In Part 2, we discuss the characteristics of network functions and operations and explore the business justifications that are helping us make the most informed decisions possible on utilizing our upgraded network.

What’s Virtualized?

We will look at activities capable of being virtualized using general-purpose, Commercial Off-The-Shelf (COTS) computers and then the functions that must remain fixed on dedicated hardware.

Hardware optimization technologies will remain effective for some time. In the domain of packet forwarding, Application-Specific Integrated Circuit (ASIC)-based performance improvements are well known and will improve further as semiconductor technology advances. Therefore, for the near term, COTS technology cannot and will not match ASIC-based performance.

Functions that require high packet performance in the data plane or low power consumption per bits transferred will continue to rely on ASIC-based components. Future innovations with COTS hardware are expected to bring improvements in the domains of Protocol Oblivious Forwarding (POF) and the new P4 language to further reduce performance gaps between custom ASICs and general-purpose machinery. The key question is whether a platform can provide an acceptable level of performance.

Conversely, network functions that require ease of integration or have high performance requirements and a shorter deployment cycle will be considered for virtual network functions operating on COTS hardware.

Physical and Virtual Functions Matrix for Telecom Domains

Business Benefits

Etisalat conducted a comprehensive study of virtualization use-case benefits from three angles: 1) new revenue generation, 2) efficiency gains (architecture and operations), and 3) cost savings.

Virtual Customer Premises Equipment (vCPE) is the use case most targeted for revenue generation, whereas virtual Optical Line Termination (vOLT) and the virtual Broadband Network Gateway (vBNG) are the most targeted for improving efficiency.

  • Enterprise Focus: SDN and software-defined data centers form a new enterprise service market known as the virtualized Data Center (vDC), which will allow enterprises to self-service their network and IT needs. The virtualized network and IT platform can be combined with enterprise data centers to build truly scalable data processing platforms with on-demand capabilities.

vCPE will allow Etisalat to transform its networking services to maximize efficiency, promote branding, reduce deployment costs, and increase telecom profits.

  • Cloud Data Center Foundation: In addition to being a tier-one use case, the vDC has potential for generating revenue by allowing enterprises to form hybrid private clouds with distributed data centers.
  • SDN + CDN: SDN can increase revenue and save costs. A Content Delivery Network (CDN) is essential for new services such as media, entertainment, and 4K/8K videos.
  • NFV: NFV will allow operators to sustain growth by utilizing a virtualized platform without increasing operational costs.

Operations

‘Content+ Services’ is a future requirement that is already profoundly changing the landscape for operators focused on efficiency, cost reduction, and Time-To-Market (TTM). This requires a simplified Operational Support System (OSS) with an open and flexible architecture able to support the model of having standard APIs and using a Development Operations (DevOps) model. The new OSS must support the self-care needs of users and manage real-time interactions between users, partners, networks, OSS, and Business Support Systems (BSSs). To avoid new management silos, management should be centralized across physical, virtual, and cloud resources.

SDN/NFV Impact on OSS

Virtualizing SDN/NFV network management brings new challenges to the OSS:

  • Support of dynamic, policy-driven, near-real-time processes
  • Mapping and tracking of virtual and physical resources
  • Configuration, capacity on demand, and scaling
  • Service chaining
  • Root-cause and customer-impact analysis
  • Tightly coupled assurance and fulfillment
  • Intuitive and user-friendly service creation
  • Integrated analytics based on real-time traffic measurements, customer behavior, and customer use
  • Service exposure via open APIs and creation of an open business partner ecosystem
  • Automated operations for both physical and virtual resources

Because physical-to-virtual mappings change in real time, troubleshooting can be a challenge. Real-time mapping tools ensure the visibility of account service chains and impact of full or partial movement of Virtual Network Functions (VNFs) across data centers. The OSS must detect malfunctioning orchestrators and automatically take corrective actions without impacting service delivery.

OSS Architecture

To benefit from virtualized networks, operational flexibility, and service agility, the current OSS must change to support End-to-End (E2E) orchestration of multi-domain, multi-vendor networks and gain in capacity to manage both physical and virtualized networks.

This requires two layers: A BSS service layer and a service orchestration layer to communicate with multiple, domain-specific orchestrators.

The BSS and service orchestration layers must be separated to manage changes and network analytics support.

The following functional architecture for the new OSS is likely to emerge:

  • The Information and Communications Technology Orchestrator (ICTO) and the Network Services Orchestrator (NSO) will configure and deliver SDN functions and services.
  • The SDN controller will interact directly with SDN-enabled devices.
  • Enterprise Mobility Suites (EMSs) and VNF Managers (VNFM) will manage physical and virtualized resources.
  • The cloud infrastructure will be managed by a cloud infrastructure manager, such as OpenStack.
  • Content will be managed by a content management system, such as OpenCMS.

During the SDN 2.0 phase, the NSO will be needed for service agility, and a hierarchical orchestrator model will emerge soon after.

Eventually, the two OSS layers will merge to form a unified ICTO; however, they may co-exist for some time into the near future.

The hierarchical orchestration model is needed to differentiate the functions of each layer, while also hiding the complexities within each domain. NSO will be responsible for service definition and service policy management. It will, in turn, push policies to the SDN controller by communicating with NFV Management and Network Orchestrators (MANOs). MANOs will manage all aspects of the virtualized network.

This architecture supports a shared model between fulfillment and service assurance, and the model is designed to update in near-real time to capture network changes.

With the help of a policy engine, physical and virtualized infrastructure resources and the EMS analytics engine will use real-time feeds to extract intelligence and take corrective actions.

This architecture allows software to be designed in domain-specific components that support parallel work by multiple teams. Eventually, all OSS applications will be built with modular components that share common capabilities, which can be customized to allow rapid assembly of new behaviors. Each component is self-contained, including data, configurable metadata, workflow behaviors, and functionality.

Evolving Role of the OSS

The orchestrator is critical to delivering the real OSS value and is supported by:

  • E2E provisioning
  • E2E resource management
  • Service models driven by business metrics

Orchestration is the heart of automating Future Mode Operations (FMOs) and service lifecycle management. With DC processing capabilities available to all connected locations, the orchestration engine will coordinate SDN activation across the transport and data center network layers to form a virtualized, distributed network.

The orchestrator will deploy services by creating virtualized data centers, integrate with a design tool so appropriate services can be described and modeled inside the service catalog, and read the catalog and deploy the service.

Service inventory is shared between fulfillment and assurance functions, providing real-time mapping of VNFs to VMs. The event manager will correlate events to affected services. The host OSS system will see the entire network. The tenant OSS, interacting with the host OSS, will provide fulfillment and assurance visibility to partners for their portion of the physical and virtualized resources.

With the multitude of applications and domains, control cannot reside in a single monolithic software orchestrator or controller. Managing a specific complex domain requires expertly designed tiered architecture with distributed NSOs sharing a global view of the services and network provided by the ICTO. The NSO houses the network intelligence necessary for its contained domains.

The Role of Analytics in Operations

Etisalat plans to build and monetize a Big Data platform using our ecosystem to capture data related to interaction from customers, enterprises, OTT players, and service developers. Operational logs generated by virtualized services and the OSSs managing the virtualized environment are key sources of analytics data.

The network must be instrumented in the data plane to track bulk and application traffic flow patterns and collect data. The analytics engine can be deployed at the edge for time-sensitive customer information and localized computation. This is useful for detecting multiple dropped calls for a mobile customer traveling to a different country, for example. The engine can also be deployed at a central place to mine customer sentiments about a CSP from data captured from multiple social sites.

Generating actionable events from collected data and acting automatically on those events to improve customer experience or SLA are essential.

By applying data extraction and mining techniques to collected data, real-time triggers are generated. The policy engine will then send directions back into the orchestration layers to self-regulate, self-optimize, and relay important network status changes.

A closed loop between assurance and orchestration, along with policy engine and analysis capabilities, will allow linking of operations with control, which will enable network self-healing.

Service Management

The enhanced Telecom Operations Map (eTOM) is a business process and information framework that is maintained by the TM Forum to support service providers in the telecom and entertainment industries. The eTOM is structured around four major lifecycle processes that exist within each participating organization:

  • Customer Relationship Management
  • Service Management and Operations
  • Resource Management and Operations
  • Supplier/Partner Relationship Management

SDN/NFV-based OSS orchestration will allow entire service and resource lifecycles to be described as an information model able to render custom views of virtualized partner infrastructures to support their customers’ networks.

FMO orchestration technology for software-defined services and service lifecycle management can be completely automated based on design-mode inputs.

Reduced TTM

TTM is guided by two key OSS management processes in the new environment:

  • Product lifecycles are guided by marketing: Global services are presented to each local market, with cost savings coming from the integration of social media and Big Data analytics.
  • Service lifecycles are directed by engineering: Orchestration relies on the ICT architecture to reduce the time requirements for service creation, deployment, and problem resolution.

Total TTM reduction is projected to reach as high as 70 percent.

Conclusion

SDN/NFV-driven transformations based on the architectural guidelines detailed in Parts 1 and 2 of this summary have been devised to dramatically reduce the time required to launch and market new services.

In line with the Etisalat vision of the network and operations landscape for 2020, the following takeaways are valuable across the industry:

  • Reducing TTM and deployment cycles is critical.
  • Operators must develop a comprehensive understanding of SDN/NFV technology and its full impact on existing networks, operations, and organizations.
  • Business cases with realistic targets need be arranged and followed to achieve a practical roadmap for SDN/NFV deployment.
  • Current investments must be future-proof in alignment with a gradual plan for adopting an SDN/NFV infrastructure.
  • Service orchestration is critical for operational agility and efficiency. Operators must insist on accelerating the development of open and unified standards for cloud Operating Systems (OSs).
  • Open architectures are fundamental for the smooth deployment and ultimate success of SDN/NFV. To achieve this outcome, operators must begin to enforce certifications, lifecycle quality control, and rigorous integration testing.
  • Vertical integration of cloud components is a risk factor for quick deployment. This can be avoided with pre-deployment integration tests and stringent verification processes.
  • System integrators with IT and telco expertise are vital for the swift deployment of cloud-based networks.

To enable new communications, Content+, and IoT services, Etisalat will ensure rapid technology innovation by our team and our partners, as we plan to implement a modular, pragmatic, and simplified approach toward the goal of maximizing network virtualization.

About Etisalat

Headquartered in Abu Dhabi, Etisalat provides innovative telecommunications solutions and services to 167 million subscribers in 18 countries across the Middle East, Asia, and Africa.

This white paper is prepared in collaboration with Huawei. For more information, please write to egwhitepaper@etisalat.ae

Copyright 2015 © Etisalat. All Rights Reserved

By Hatem Bamatraf

Chief Technology Officer, Etisalat Group