As the global leader for OpenStack at Red Hat, Balakrishnan is responsible for driving business strategy, product management, partner strategy and worldwide gotomarket strategy across enterprise & Telco/NFV segments for Red Hat Enterprise Linux OpenStack Platform .
Balakrishnan has 20 years of experience in the IT industry across various enterprise technologies. Prior to joining Red Hat, Balakrishnan was with Microsoft where he held various product management and product marketing roles across Azure, System Center, Windows Server,Windows Storage Server, Exchange Server and Windows desktop products and technologies.
SDxCentral: How are service providers adopting NFV?
Balakrishnan: Network Functions Virtualization (NFV) is a key initiative being undertaken by many of the Tier 1 carriers across the globe to react to the changing competitive and economic realities faced by them. NFV aims to address two key challenges bringing costs in line with the revenue growth and to improve ‘service velocity’ or the agility with which services can be made available. Some notable carriers that have implemented these NFV solutions include: Nokia, Ericsson, AT&T, Telefonica and others. Compared to the traditional approach of proprietary hardware and software, NFV is predicated on leveraging industry standard servers, switches, and storage to reduce the cost and increase the speed of service delivery for fixed and mobile networking functions. OpenStack has become the lingua franca of NFV infrastructure (NFVi) layer of the solution with hundreds of proof of concepts underway across the globe,fueled by the backing of industry players to forums such as OPNFV. Service providers across the globe are conducting trials or proof of concepts to validate the technology, prioritize use cases and select vendors to partner with in their journey to driving agility and reducing costs via NFV. We’re also seeing some of the service providers rolling out initial deployments to test/prove the operational management of the new infrastructure with an intent to rollout new services in production in the second half of the calendar year. We expect to see more of these trials turning into production in the coming year, fueled by the success of the early adopters and maturation of technology as well as the ecosystem
Does the service provider community have specific needs to make NFV solutions “carrier grade?”
Balakrishnan: There is some lack of clarity around carrier grade NFV requirements when it comes to NFV implementations, not helped by the fact that traditional players in this space are touting their capabilities as carrier grade NFV.
While the NFV infrastructure landscape is evolving, carriers are faced with the reality that high availability and service level agreements have been fundamental requirements to their business. In the traditional approach, carriers have been living by/delivering Five9’s reliability through proprietary hardware and software as well as redundancy at multiple layers. With NFV, the approach to high availability of virtualized network functions (VNFs) changes from an active standby or active deployment to a virtualized application using the inherent services and capabilities of the underlying NFVi layer. With this approach, whenever there is a failure, the impacted traffic will be redirected to a new instance or a load shared instance of that application.
Another dimension of need is around performance i.e. to make network processing in virtual machines as fast as bare metal x86 machines. This has been recently achieved and proved using multiple approaches. The current goal is to improve the x86 virtualized systems even more and perform better than purpose made network equipment, thanks to concurrent usage of multi-core CPUs.
The carrier grade NFV needs differ from one application to the next, which is why it is really important to collaborate through the entire stack, including hardware, software platform, and network application vendors to help ensure that a complete NFV solution can meet carrier needs.
What kind of enhancements are needed to make OpenStack carrier grade?
Balakrishnan: It’s important to realize that to make an NFV implementation successful or to meet the needs of carriers, all layers of the stack need i.e. hardware, NFVi (OpenStack), VNF and MANO layers need to be architected to meet the needs. In other words, there are capabilities required at all layers including OpenStack to meet the service level agreements requirements.
Within that context, there are areas in Openstack such as high availability and security there is opportunity to further enhance the capabilities to meet the operational management requirements if an NFV implementation.
How can service providers and cloud providers work with the open source community to produce carrier grade solutions?
Balakrishnan: As a pure open source company, Red Hat‘s core value proposition revolves around being the catalyst for open innovation and in this case, to work with service providers and the network equipment providers (NEP) to identify, prioritize and partner to develop capabilities in a truly upstream aligned manner. In addition, industry forums such as the OPNFV community provide a forum to drive this in an aligned manner.
This collaboration creates the opportunity for the industry to work together to get baseline performance and reliability for the open source components which make up the platform, and to define the set of interfaces which applications need, and which management platform vendors will be able to use to manage the lifecycle and quality of service of network functions.
By participating in the OPNFV project, representative cross section of the industry can collaborate on defining requirements, and translating those requirements into feature gaps in the platform. We have taken a leadership role in helping the OPNFV community implement those gaps, and get those changes upstream to the open source projects, allowing them to benefit everybody.
Is it important to have a broad number of partnerships to attain these goals?
Balakrishnan: Absolutely an NFV deployment includes physical infrastructure, software infrastructure, management applications, and multiple applications. It would be difficult for a single vendor to provide all that, and the integration services needed to bring these disparate components together. By building a broad ecosystem of partners, Red Hat can enable service providers to get an end to end supported NFV solution, and to deploy with greater confidence.
What’s the future of open source in the service provider community?
Balakrishnan: Open source has gained importance in the datacenter, and all indications are that in the immediate future the service provider community will be no different. NFV platforms can build on Linux, and can include open source components for data plane acceleration, virtual switching, software-defined networking and infrastructure as a service. Network functions can run on Linux guests, with an optimized userspace to seek to provide high performance. And we are seeing a trend towards open source management applications and components to manage the lifecycle of those applications.
We do not expect to see significant penetration for open source network functions in the near term. However many ISVs use open source tools and libraries to build their solutions, which can mean that it’s a Linux kernel running the VNF. We believe the future can have an ecosystem of certified VNFs and NFVi platforms with Linux as the common denominator. On the other hand, we also believe that applications such as OSS, BSS, and element management systems, which include a great deal of domain specific information, will also remain dominated by proprietary vendors in the near term.