Skip navigation
2016

The Wind River Titanium Server network virtualization platform was recently selected by LGU+ for their first commercial carrier grade Network Functions Virtualization (NFV) deployment.

 

Post1.pngFormerly known as LG Telecom, LGU+ is one of the 3 major operators in South Korea and owned by LG Corporation. LGU+ provides a wide range of wired and wireless communications services such as triple-play and data services. They were one of the first operators worldwide to introduce an unlimited LTE data plan.

 

We spoke recently with Mr. Jaeho Choi, the General Manager and Project Leader responsible for the virtual Router program at LGU+. In this post we’ll summarize what we learned from Mr. Choi about the business challenges that are driving the move to network virtualization, why they selected Titanium Server for this project and their plans for future developments in this area.

 

Virtual Router as NFV test case

 

During 2015, LGU+ conducted a number of studies to determine the optimum use case for their first application of carrier grade NFV. Besides the expected ROI, a key criterion was to minimize disruptions to their existing operational processes while also minimizing the schedule risk for this initial project. After exploring various potential use cases , the team settled on the virtual Router as most appropriate for their commercial carrier grade NFV program in 2016.

 

Service uptime is critical

 

High availability and reliability for field deployment was top of mind for LGU+. The operations team at LGU+ has an excellent track record of maintaining the service uptime that their enterprise customers expect and that is required for critical services. For them, a key concern when introducing NFV into their network is ensuring that the new virtualized services will meet or exceed current uptime levels. Mr. Choi stressed that this constraint was a key factor in their selection of virtualization platform.

 

Flexibility and open solutions

 

Finding a solution with the flexibility to scale and quickly deploy new services as needed was an essential consideration for LGU+. As such, “open” solutions were also an absolute requirement for LGU+. Recognizing that a large number of vendors are introducing innovative new hardware and software for NFV on a regular basis, they want to ensure that they retain the flexibility to select the optimum products for their applications. It’s critical to avoid any risk of vendor lock-in, so they only considered products that are fully compatible with all applicable open standards and that have demonstrated interoperability with complementary products from other vendors.

 

CAPEX and OPEX reductions driven by network virtualization

 

LGU+ sees the potential for significant reductions in both CAPEX and OPEX from their virtual Router program. Migrating to a virtualized implementation that leverages the performance acceleration features of Titanium Server enables them to implement the same function on just two high-performance servers, which leaks to significant cost savings while also delivering other benefits such as improved scalability and better resource utilization.

 

Why Titanium Server?

 

Mr. Choi outlined five main reasons for selecting Titanium Server for the virtual Router program at LGU+.

 

Post2.pngFirst, the performance acceleration achieved by the Titanium Server Accelerated vSwitch (AVS) maximizes Virtual Machine (VM) density, which in turn maximizes the number of virtual Router instances running on a single server. Typically, Virtual Network Functions (VNFs) based on the Intel® DPDK library deliver up to a 40x performance improvement when running on Titanium Server compared to Open vSwitch (OVS), while those using standard Linux kernel I/O realize up to 9x acceleration. In each case, this is achieved with no change to the VNF code itself, leveraging open-source Accelerated Virtual Port (AVP) drivers to deliver the performance gains.

 

Second, Titanium Server ensures six-nines availability at the infrastructure level, enabling LGU+ to meet their service uptime requirements and be in compliance with all their applicable Service Level Agreements (SLAs).

 

This Carrier Grade reliability is achieved through a comprehensive suite of features and capabilities, including: accelerated live VM migration; the live migration of DPDK-based VMs; accelerated detection of and recovery from both host and guest failures; deterministic, low-latency interrupt latency; comprehensive telco-grade AAA security (Authentication, Authorization and Accounting); automated VM scaling; accelerated virtual switching.

 

Third, Titanium Server is a fully open solution, proven to be compatible with all the relevant open standards and APIs. As a major contributor to open-source communities, Wind River upstreams patches to OpenStack and KVM that address critical needs for telecom customers. As an example, Wind River is ranked in the top 10% of contributors to the Nova Compute project, which is the largest project in the OpenStack community. Titanium Server eliminates the risk of vendor lock-in through guaranteed interoperability and interchangeability.

 

Fourth, as a fully-integrated, packaged NFV cloud Titanium Server is available for immediate deployment into live networks. Its advanced installation, commissioning and upgrade capabilities not only accelerate the initial deployment process but also streamline subsequent seamless updates and patching.

 

Finally, LGU+ sees great value in the extensive range of products from industry-leading partners that have been validated through the Titanium Cloud ecosystem. This ecosystem currently comprises 37 partners with additional companies joining regularly. LGU+ welcomes mutual certification among ecosystem vendors in order to optimize go-to-market time, improve efficiency and reduce their schedule risk.

 

Next steps build on initial successes

 

LG U+ plans to complete the benchmark test stage of their virtual router project, deploy it in their live network for field tests and then for commercial service.

 

After the LG U+ operations team has completed satisfactory stability and performance testing of the virtual Router in their live network, they plan to deploy additional instances and the virtual Router will be deployed as a backup for physical appliances. In the future, migration of their complete virtual Router infrastructure from physical to virtual implementations will be considered.

 

Our work with LGU+ has already delivered positive results for their virtual Router use case we’re very grateful for the opportunity to collaborate with Mr. Choi and his team. We can’t wait to get started helping them take the next steps in their NFV journey.

China Unicom recently selected Wind River’s Titanium Server network virtualization platform for an in-depth evaluation as part of their wide-ranging strategy for Network Functions Virtualization (NFV). This selection offers interesting insights into the business opportunities and technical priorities that are important to China Unicom as they start to migrate from physical to virtual infrastructure.

 

Post1.pngChina United Network Communications Group Co. Ltd ("China Unicom") was formed in 2009 after the merger between the former China Netcom and the former China Unicom. It has operations in 31 provinces across China as well as in many other countries around the world.

 

China Unicom operates services that include: fixed and mobile communications; domestic and international communications; satellite International Private Leased Circuit (IPLC); data communications; network access; value-added telecom; system integration. In 2012, China Unicom became a founding member of the Cloud Computing Industry Alliance in Beijing, other members of which include companies like Alibaba, Baidu and Tencent.

 

Within China Unicom’s Network Technology Research Institute, Mr. Tong Junjie is the engineer responsible for the selection of Titanium Server, as part of his overall role in NFV technologies and solutions. In this post, we’ll summarize what he told us about the business opportunities that China Unicom sees with NFV, the technical challenges that are most important to them and their plans for future work in this area.

 

NFV for new revenues as well as cost savings

 

China Unicom is looking to NFV as an approach to achieve top-line revenue growth as well as operational cost reductions. In the near term, they expect that the service agility provided by NFV will allow them to accelerate both the trials and the deployments of new, value-added services, bringing new revenues from their vast subscriber base. In the future, they expect to also offer a range of Internet of Things (IoT) services, expanding their customer base further and thereby adding additional sources of revenue.

 

Of course NFV is also about operational cost savings and China Unicom has high expectations in that area too. They believe that replacing traditional vertically-integrated products with decoupled, multi-vendor solutions will allow them to achieve significant reductions in both CAPEX and OPEX. One key to deploying decoupled solutions is selecting software elements that are fully compatible with open standards, which is one of the reasons that they selected Titanium Server.

 

Early use cases include virtual CPE, virtual EPC and virtual IMS

 

China Unicom is focused on three initial use cases for NFV. These include virtual Customer Premise Equipment (vCPE), virtual Enhanced Packet Core (vEPC) and virtual IP Multimedia Subsystem (vIMS).

 

Post2.pngTo accelerate the deployment of all these use cases, they can leverage the extensive range of products from industry-leading partners that have been validated through Wind River’s Titanium Cloud ecosystem. This allows them to select Virtual Network Functions (VNFs), service orchestrators and hardware platforms, including several from local Chinese companies, which have already been validated to demonstrate full interoperability with Titanium Server. The validation work that Wind River has already completed with these partners removes much of the schedule risk associated with China Unicom’s own, in-house testing of the multi-vendor configurations that they select for these use cases.

 

Service uptime is critical

 

China Unicom is proud of the ability of today’s physical infrastructure to maintain the service uptime that their enterprise customers expect and that is required for critical services. As they introduce NFV into their network, they need to ensure that the new virtualized services will meet or exceed current uptime levels.

 

Mr. Junjie stressed that this requirement was a key factor in their selection of Titanium Server. By guaranteeing six-nines availability at the infrastructure level, Titanium Server enables China Unicom to meet their service uptime requirements and meet their customers’ Service Level Agreements (SLAs).

 

Performance is key to OPEX savings

 

With OPEX savings being a major objective for China Unicom as they introduce virtualized services, one of the most important metrics is the number of subscribers that they can support on a single server. The more subscribers per blade, the lower the cost-per-subscriber.

 

As Mr. Junjie explored various network virtualization platforms, he realized that Titanium Server provides unique and compelling benefits in this area. Post4.png

 

With its Accelerated Virtual Switch (AVS) that delivers up to 40x the switching performance of Open vSwitch (OVS), Titanium Server minimizes the number of processor cores that are required to run the virtual switch, maximizing the number of cores available for running Virtual Machines (VMs). This in turn maximizes the number of VMs that can run on a single server, thereby also maximizing the number of subscribers per server and minimizing the critical cost-per-subscriber.

 

Through the Titanium Cloud ecosystem, Wind River works with its VNF partners to ensure that their VNFs fully leverage this performance acceleration.

 

VNFs that are not based on the Intel® DPDK library typically achieve up to a 9x performance improvement when using AVS together with the open-source Accelerated Virtual Port (AVP) Kernel Loadable Module (KLM). Those based on DPDK see up to a 40x improvement thanks to the open-source DPDK Poll-Mode Driver (PMD). In both cases, no changes are required to the VNF code itself (more details here).

 

Open solutions are mandatory

 

For China Unicom, “openness” means compliance with open standards, which leads to interoperability at all levels of the architecture and removes the risk of vendor lock-in. Mr. Junjie explained that these are mandatory elements of NFV solutions because they will enable China Unicom to sustain the business benefits that they expect to achieve from NFV.

 

Titanium Server is a fully open solution, proven to be compatible with all the relevant open standards and APIs. As a major contributor to open-source communities, Wind River upstreams patches to OpenStack and KVM that address critical needs for telecom customers. The rigorous interoperability testing performed with partner products through the Titanium Cloud ecosystem provides tangible proof of compatibility with open standards and APIs.

 

Looking forward to future collaboration

 

China Mobile’s comprehensive testing of Titanium Server is progressing well.

 

As this phase of the program completes, we look forward to helping Mr. Junjie and his colleagues to deploy vCPE, vEPC and vIMS use cases in their networks. All indications are that they’ll be able to achieve both the new revenue and the cost savings that they’re looking for.

Fog computing is an emerging trend that distributes computing for more timely analytics. The OpenFog Consortium — founded in 2015 by Intel and other industry leaders — is helping define fog computing architectures and ensure interoperability. See the strides the Consortium is making, especially around system architecture specifications and IoT gateways that localize data analysis.

 

Read this article on insight.tech>>

Keeping power consumption in check is a serious challenge when you need deep intelligence. The Intel® Quark™ SE microcontroller C1000 tackles this issue with an innovative pattern-matching engine and smart sensor subsystem. Discover the benefits of offloading processing from the main core to deliver edge intelligence at low power.

 

Read this article on insight.tech>>

Few companies have expertise in both embedded design and Big Data — or know how to combine these domains. The Advantech IoT Gateway Starter Kit simplifies matters. Find out how this gateway comes preloaded with cloud connectors and a drag-and-drop interface that powers sophisticated data analytics.

 

Read this article on insight.tech>>

Filter Blog

By date: By tag: