xFlow Research

Network Infrastructure Services

NFV Infrastructure

Design and development of compute, network, storage, and control resources to build scalable, high-performance NFV infrastructure.

NFV Infrastructure

xFlow Research provides specialized services in NFV Infrastructure (NFVI), enabling efficient, scalable environments for deploying Virtual Network Functions (VNFs). We design and integrate compute, control, network, and storage resources to deliver robust, high-performance NFV solutions.

Why NFVI?

  • End-to-end expertise – from physical network/storage design to overlay networks
  • Performance optimization – NUMA awareness, CPU pinning, HugePages provisioning
  • Automation-first – rapid PoCs and demos using OpenStack, Ansible, Puppet, or Chef
  • High throughput – leveraging DPDK, OVS-DPDK, and SR-IOV for near line-rate packet processing

Our Expertise

  • Physical & virtual network/storage architecture
  • Overlay network design & resource optimization
  • Automated provisioning (HugePages, NUMA, CPU pinning)
  • OpenStack integration & custom extensions for SR-IOV and VNF cluster management
  • Development of DPDK-based applications (e.g., GTP testers, echo servers, traffic generators)
  • Accelerated packet switching with OVS-DPDK
  • VNF lifecycle management via message bus (RabbitMQ) and automation frameworks (Ansible, Puppet, Chef)

DPDK – Data Plane Development Kit

We build high-performance applications using DPDK to accelerate packet processing on x86 platforms. Our expertise covers:

  • Development of telecom testers (GTP-U, GTP-C), echo servers, traffic generators, and proxy applications for vEPC clusters
  • Optimizing VNFs to achieve near line-rate throughput and reduced latency in virtualized environments
  • Integrating DPDK into custom applications for performance-critical telecom and enterprise workloads

 

OVS-DPDK – Accelerated Open vSwitch

We specialize in OVS-DPDK, a high-performance version of Open vSwitch enhanced with Intel® DPDK libraries for efficient packet forwarding. Our work includes:

  • Deploying and managing OVS-DPDK in NFVI environments
  • Integration with OpenStack for flexible orchestration
  • Enabling accelerated switching to meet carrier-grade performance requirements

OpenStack

xFlow Research leverages OpenStack to manage large pools of compute, storage, and networking resources in NFV data centers. Our contributions include:

  • Automated deployment of OpenStack for quick PoCs and demos
  • Development of extensions to manage VNF clusters and SR-IOV NICs
  • Nova extensions for automated SR-IOV provisioning (CLI and GUI support, compatible with OpenStack Havana & RHEL 6.5)
  • Dashboard integration for simplified VNF lifecycle management

SR-IOV – Single Root I/O Virtualization

We enable performance-centric NFV and cloud solutions using Intel’s SR-IOV technology. Our capabilities include:

  • Automated SR-IOV provisioning via OpenStack Nova extensions
  • Assigning SR-IOV NICs through CLI and GUI interfaces
  • Delivering virtualization solutions that balance flexibility with near-native performance

Message Bus

We use message bus systems such as RabbitMQ to manage VNF lifecycles. Our implementations support:

  • Sending control messages/commands between orchestration systems and VMs
  • Instantiating, configuring, modifying, and tearing down VNFs (VNFM)
  • Enabling scalable, event-driven VNF management frameworks

Ansible

We harness Ansible to automate VNF lifecycle management and NFVI configuration. Our work includes:

  • Writing playbooks to instantiate, configure, and manage VNFs
  • Automating VNF provisioning, updates, and teardown
  • Supporting Puppet and Chef for flexible automation options