15 Jul 2015
NFV Technology & Transformation of Network Architectures
Glenn Seiler, Vice President, Software Defined Networks, Wind River looks at the way NFV - Network Functions Virtualization is transforming network architectures.
Explosive growth in data traffic driven by the demand for video, the increasing use of mobile devices and cloud-based services is overwhelming the capabilities of today’s networks.
The challenges for service providers are to handle this massively increased demand for bandwidth while also reducing capital and operating expenses and deploying new services to generate new revenue streams and remain competitive with the major ‘over-the-top’ content providers such as Google and Amazon.
The network service providers and telecommunication carriers are increasingly looking at technologies such as network functions virtualization, NFV to modernize and transform their networks to achieve these goals.
NFV & network transformation
This transformation will require new architectures that are agile enough to rapidly respond to changing traffic patterns, new applications and fast evolving user requirements. It is the implementation of NFV technologies that can enable network service providers and telecommunication carriers to introduce new services faster, while reducing the costs of building up network capacity and controlling day-to-day operational expenses.
The primary concept of NFV is to transform existing multiple network functions or services into software applications, then consolidate them and run them on virtual machines, VMs, based on commercial-off-the-shelf (COTS) hardware. Historically the telecommunications network infrastructure has been based upon proprietary software running on dedicated single-function and purpose-built hardware.
Now network providers and carriers are looking at the economies of scale and standardization offered by the IT industry including virtualization, standardized servers and operating systems, and open-source software such as Linux and the OpenStack cloud computing platform for public and private clouds, and to bring these technologies to the carrier network. However, NFV implementation will require carrier-specific virtualization software that can deliver real-time traffic capability and the scalability to handle billions of newly connected devices and the data traffic generated by each of these while supporting high reliability and availability to deliver the necessary minimum five nines (99.999%) or six nines (99.9999%) of uptime.
The network transformation is already happening with widespread NFV proof of concepts (PoCs) and trials starting, in conjunction with the closely associated technology of SDN (software defined networking), and the emergence of cloud services. Major telecom carriers and network service providers are already collaborating in the ETSI Industry Specification Group for NFV (NFV ISG) to drive the necessary specifications and architecture to enable flexible and scalable network architectures that will allow the quick deployment of new services, while also reduce operating costs via efficiencies in hardware utilization and energy consumption.
However, significant challenges remain in this transition to NFV-based network architectures that fundamentally cannot be met by traditional IT-based solutions. Firstly, management models will be required for provisioning and configuring services: operators will need to manage and integrate new management systems with their existing OSS tools and the network element management tools that handle network devices. Secondly they will need to maintain a highly available and reliable 5 9s or 6 9s system to ensure current levels of SLAs (Service Level Agreements).
The third challenge is performance, as there can likely be a significant degradation due to virtualization on standardized commodity hardware. Solutions today are highly vertical with a highly optimized piece of hardware for tasks such as a gateways, evolved packet core or packet inspection device, and typically these will have an element management tool used for each device. The goal of NFV is to take these and consolidate them onto a single piece of common hardware. However, this could mean a reduction in the levels of optimization and data performance in a virtualized environment on what used to be optimized hardware.
NFV & requirements
The major requirements for delivering carrier-grade communications certainly include high reliability and the five or preferably six nines of network uptime availability, but also need to address all aspects of the system including management, performance, scalability and security.
Fundamentally, traditional IT solutions are not designed to meet carrier grade requirements: an example, enhancements will need to be made to OpenStack to support the live migration of a virtual machine from one host to another within the very short times required for telecom services. Migration will need to be done in seconds rather than many minutes. Another requirement is fast detection of single or multiple software/hardware faults that can cause the failure of a virtual machine. Fault detection and automatic restart needs to be within milliseconds to achieve minimal loss of service or data, whereas in a standard OpenStack implementation it could take several minutes including a manual restart.
In terms of performance, there are many different types of devices, many of which such as baseband radio will require high-performance and very low levels of latency. A high-performance, highly reliable and secure hypervisor is the first step for NFV, but performance, reliability and security need to be applied to all aspects of an NFV host or server. One critical aspect of performance is the communications between applications or services residing on a node. A high-performance virtual switch (vSwitch) is critical to manage the flow of communications to and between the virtualized network functions (VNFs) running on a server.
Further requirements are the necessity of highly scalable solutions to handle hundreds of thousands of subscribers and highly diverging levels of data demand, and finally a high level of security will be required including encrypted AAA (authentication, authorization and accounting) databases, network level authentication and encryption services.
Flexible NFV solutions
One of the benefits of NFV is that Service Providers and telecommunications carriers can now consider the possibilities of building solutions themselves from standard enterprise IT products and other readily available open-source technologies. However, the development of a complete NFV platform can cost tens of millions of dollars and take perhaps 12 to 18 months to develop a mature solution that achieves the necessary reliability and carrier grade network availability.
Wind River offers a wide portfolio of products that address the network transformation, including many Open Source based products that are suitable for companies that want to build their own solution. These include Linux-based virtualization and OpenStack.
The Wind River component products include Wind River Linux, which was developed from the Yocto Project development infrastructure and is a commercial grade Linux platform that offers a broad set of services with a fully integrated build system and development tools. An option for Wind River Linux is Open Virtualization, which optimizes open-source Kernel-based Virtual Machine (KVM) technology to allow the deployment of network services on virtual machines without the performance loss associated with using traditional IT virtualization products. It meets carrier grade requirements and delivers near-native hardware performance speeds, whereas standard KVM maximum latency is typically several thousand percent higher than native results.
However, combining a broad range of NFV-based applications across the network stack can be a costly and time-consuming stage for carriers and service providers. The Wind River Titanium Server (block diagram is shown in figure 1) is a fully integrated solution that is based on Open Virtualization with Wind River Linux, optimized real-time Kernel-Based Virtual Machine (KVM++), carrier grade enhancements to OpenStack, and also the Data Plane Development Kit (DPDK) accelerated vSwitch technologies. The tool delivers a turnkey platform for carrier-grade virtualization and significantly reduces time-to-market and operational expenses for customers, accelerating the path to NFV. Unlike existing enterprise solutions, Wind River Titanium Server is a commercial carrier grade software platform for NFV delivering the six 9s reliability that is required for the telecom industry.
Figure 1 – Wind River Titanium Server
Network providers and carriers face significant challenges to develop new revenue generating services while also expanding their network capacity to cope with the almost exponential growth in demand for bandwidth.
These companies will need to leverage traditional IT and open-source technologies, to build high-performance NFV solutions that deliver robust low-latency virtualization, while also supporting the reliability and availability requirements for un-compromised carrier grade performance along with real service agility and application scalability.
The implementation of NFV technologies via easy to deploy solutions will enable them to transform their networks to simultaneously introduce new services faster and reduce their costs of building new network capacity and make network operations easier and less expensive.
Page 1 of 1
About the author
Glenn Seiler is vice president of product management at Wind River. Glenn is responsible for creating and managing Networking product initiatives that solve critical problems for Wind River customers developing solutions for network functions virtualization (NFV) and the network transformation. Glenn has been managing platform software for more than 30 years and has been a strong advocate for Open Source and Open Standards. Glenn was a founding member of the Carrier Grade Linux Workgroup, was a contributor to the SCOPE Alliance Carrier Grade OS committee and was the Vice President of the OpenSAF Foundation. Glenn has been with Wind River since 2005, and prior to Wind River Glenn held Director of Product Management positions at MontaVista Software and BSDi and Unix Product Management roles at SCO and Texas Instruments. Glenn holds a B.S. in computer science from Stephen F. Austin University and a M.S. in computer science from Texas State University.
Wind River Systems, Inc. is a wholly owned subsidiary of Intel and it is a provider of embedded systems technology, development tools for embedded systems, middleware, and other types of software accompanied by an award winning level of customer support. The company was founded in Berkeley, California in 1981 by Jerry Fiddler and David Wilner.
Most popular articles in Telecoms & networks
Share this page
Want more like this? Register for our newsletter