Network Functions Virtualization – Building the Software Telco

Network Functions Virtualization – Building the Software Telco

View the Q&A

 

Join Tom Schroer, Director of Service Provider Marketing with Dialogic and James Middleton of Telecoms.com as they discuss what organizations need to know about Network Functions Virtualization and strategies to transform today’s service delivery networks to tomorrow’s Software Telcos.

Service providers are under pressure to constantly stay at the forefront of technology, delivering new and innovative services to their customers. To keep pace with these demands, they find themselves caught in a rapid and continuous deployment lifecycle reliant on purpose-built hardware appliances. But this is a strategy that is getting harder to manage and runs the risk of slowing innovation and increasing operational costs.

To address these challenges, Service Providers are driving a transformative concept called Network Functions Virtualization – or NFV – that moves network functionality to software and leverages commercially available commodity server hardware from the IT sector and teams it with virtualization technology.

In this webinar, we’ll discuss the following concepts

  • The NFV framework and the challenges associated with virtualizing real-time media intensive services
  • What criteria can be used to assess whether a particular network element is a good candidate for virtualization
  • Strategies service providers can use to help kick start their NFV implementation with virtualizing session control and management at the service provider and enterprise access edge
  • Followed by Q&A

“Programmable Panacea or Executable Nightmare – What Telcos Need to Know About Session Virtualization”

www.dialogic.com/nfv_whitepaper

Tags; Archive, Dialogic, Dialogic, Featured, NFV framework, Service Provider, telecoms.com
Q&A
  • Nicole Ramson
    Nicole Ramson September 24, 2013 at 9:37 am

    The Q&A is now closed. Please direct any further questions to Tom Schroer by emailing Thomas.Schroer@dialogic.com.

  • louie jamison September 24, 2013 at 9:16 am

    Will some parts of the network or specific network functions not be able to be virtualized?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:26 am

      While innovations in commercial general purpose hardware are able to help virtualize many applications like control plane and data plane packet processing or crypto there may be applications that may not scale as well in pure software environment as with specialized hardware like digital signal processing.

  • JimDev September 24, 2013 at 9:09 am

    Is the ETSI work being done primarily standards creation or is there some kind of interop being contemplated?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:25 am

      There is a lot of activity within the ISG on establishing recommendations for the end-to-end architecture, security, requirements, use cases and so forth. There is activity under way to set the groundwork for Proof of concepts and we are already seeing announcements from some of the ecosystem participants on that front.

  • billnewley September 24, 2013 at 9:08 am

    Who will be the biggest benefactor from NFV? Hardware vendors, OSS/BSS vendors or service providers?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:11 am

      I think the service providers will definitely benefit in this initiative for the reasons we covered, but I don’t think it will be limited to them. This is one initiative where there is benefit to the entire ecosystem including the customers and vendors. Because one of the outcomes of this initiative is to decouple the software applications from the underlying hardware, it opens up opportunities for additional innovation by software developers and vendors who were somewhat excluded due to the hardware intensive aspect of traditional telephony infrastructure. So I see this as a win-win for most of the participants.

  • billnewley September 24, 2013 at 9:05 am

    I know things like SR IOV and DPDK are helping to improve performance, what else can be done to accelerate performance?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:12 am

      SR IOV and DPDK are base technologies to improve network IO performance. Compute and storage resources are equally important. Pooling, effective utilization and orchestration of these resources between VMs are important as well to achieve overall performance acceleration.

  • Paul Hemming September 24, 2013 at 8:58 am

    Thanks, Tom.
    Do the network engineering, IT and marketing functions have to collaborate closely in order to make an NFV approach work effectively?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:20 am

      I think so, NFV pushes far into the realm of IT as more functionality is moved from the central office to the data center. The underlying hardware – server virtualization, management, COTS servers, etc. – have been the in the realm of the IT organizations and I believe you are seeing this merging of silos. The orchestration of disparate VNFs will be critical to take advantage of the benefits from automation. From a marketing perspective let alone business model perspective, service providers will have the ability to look at different service models as well as be an enabler for application providers that can leverage their virtualized environments to provide services to customers.

  • Christophe September 24, 2013 at 8:54 am

    Any telco which has implemented a pilot with customer?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:30 am

      The NFV community is working on framework for telco operators and infrastructure vendors to come together and do Public PoC demo to not only demo the proof of concept from technology perspective but also for whole ecosystem to come together, so we should soon see this. There are other members in the ecosystem that have announced they’re starting proof of concepts as well so you’ll start seeing more activity on this front.

  • Nicole Ramson
    Nicole Ramson September 24, 2013 at 8:54 am

    Thank you for attending this webinar. Tom will now be answering your questions.

  • Ian Ginn September 24, 2013 at 8:53 am

    Hi, is there a way of receiving a link to this recording offline?

    br,
    Ian

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 8:56 am

      Yes, it will be posted on the Informa website, you will also receive the link by email

  • Christophe September 24, 2013 at 8:44 am

    where is such a solution already implemented?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:23 am

      We just released the virtualized SBC, and we have trials ongoing with enterprise customers and carriers for the solution. This is from VNF virtual appliance perspective.

  • Pål Grønsund September 24, 2013 at 8:43 am

    What are the requirements and limitations on COTS for VNFs?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:04 am

      COTs are part of NFVI (Network function Virtualization infrastructure domain) and requires at a minimum the support of virtualization technologies that allows the abstraction and pooling of compute, storage and networking resources for sharing between virtual machines running the VNFs.

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:09 am

      A limitation could be applications that requires specialized hardware and may not be possible on COTS hardware like digital signal processing. In certain instances where the operator has centralized large amounts of media plane processing, dsps in purpose built hardware may be an approach.

  • nataliecompton September 24, 2013 at 8:32 am

    What kind of performance degradation would be seen when comparing bare metal and virtualized environments?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:01 am

      It depends on the type of workload and the virtualization technologies in use. Results from some infrastructure vendors suggest that you can achieve near bare metal performance when specific virtualization features are used.

  • louie jamison September 24, 2013 at 8:31 am

    What applications do you feel will be good candidates for virtualization?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 8:58 am

      In the webinar we talked about some criteria to use when assessing candidates for virtualization. There is considerable activity in the ISG around proof of concept use cases that range from EPC, IMS to CPE opportunities. Like we indicated, it will be a combination of functional role and location or position in the network and the underlying software resource requirements. We believe of course that session control or session border controllers are good candidates especially for the use cases we cited. Other applications would include application servers for UC or IP telephony along with voice core platforms such as in IMS or soft switches.

  • azegouar September 24, 2013 at 8:30 am

    Hello, Could you provide us the slides of the prez ?

  • JimDev September 24, 2013 at 8:29 am

    Wouldn’t a SBC be easier to think of as a decomposed set of components compared to a single application? For example, what you virtualize and then would orchestrate would be different modules that together comprise an SBC, i.e. media processing; access control lists; transcoding; header manipulation; security etc.

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 8:57 am

      Interesting approach and something that virtualization and NFV enable. Initially, we are establishing our Virtualized SBC as single VNF, however, you bring up an interesting evolution path where the different modules that make up an SBC could at some point be separated as individual VNFs that can be orchestrated as needed. We are taking that approach to an extent with our Virtualized SBC and our architecture provides a lot of flexibility and you can configure it accordingly to do specific functions and not others, for example control plane only and not intercept the media – so it lends itself to the approach of decomposing different functions. In our upcoming release we will enabling the software transcoding capability as part of our BorderNet virtualized SBC, however, we will be architecting it such that in the future, the application will be able to leverage separate transcoding resources. This opens up the possibility for a virtualized transcoding VNF as well as enabling the use of an external transcoding resource – physical or virtualized.

  • Pål Grønsund September 24, 2013 at 8:23 am

    Many network functions can be virtualized, but what do you consider as the most important and beneficial network functions to be virtualized for an operator?

    • Thomas Schroer
      Thomas Schroer September 24, 2013 at 9:02 am

      We went through several use cases where we think service providers can start seeing benefits and those are based on activity we are seeing. The service provider access edge is a good area where operators can start to virtualize infrastructure and leverage their investment in data center infrastructure that will help them enable new capabilities with their enterprise customers. The ETSI NFV community is also working on use cases that identify areas of the network that can be virtualized such as in delivering services to enterprises, residential customers, the mobile core and IMS.