NFVTime-OS

The NFVi OS that turns any x86 into a carrier class vCPE


The heart of the NFVTime is an open source, thin and robust NFVi-OS that turns any x86-device, low or high resources, to a fully operational vCPE. With a very low system footprint and design optimized to extract superior performance from Intel™ processors, it is the perfect solution for operators seeking to rapidly deploy end-to-end, carrier-class networking solution for business or metro environment. 
The heart of the NFVTime is an open source, thin and robust NFVi-OS that turns any x86-device, low or high resources, to a fully operational vCPE. With a very low system footprint and design optimized to extract superior performance from Intel™ processors, it is the perfect solution for operators seeking to rapidly deploy end-to-end, carrier-class networking solution for business or metro environment. 

NFVTime-OS allows complete decoupling between hardware, software and VNF, enabling mix and match to optimize the service provider flexibility, Cap-Ex, Op-Ex and most importantly, utilize the community innovation and quickly adopt new hardware, services and architecture versions. It can run any VNF service out of the box and requires zero setup thanks to the NFVTime-Central provisioning and control server. 

NFVTime-OS is tested and certified on many hardware types, obtained from any vendor or directly from Telco Systems.

NFVTime-OS support out of the box broad hardware from leading HW vendors and broad VNF library. With the backup of NFVTime-central, it automate and allow swift deployment of NFV services in mettar of weeks from decision to service luanch.
NFVTime-OS Architecture
NFVTime-OS architecture is based on open source and community projects optimized to achieve carrier class vCPE. NFVTime uses RT Kernel to optimize and improve performance of the overall vCPE and VNFs performance.

The major NFVTime-OS building blocks include:
Open vSwitch (OVS) is used as the virtual networking component. It connects VNFs' virtual ports between themselves and to the external ports. It also separates management and data traffic on the same WAN uplink port. 
Telco Systems solution support both VirtIP and DPDK and SRIOV to  improved packet processing performance

Selected OpenStack components:

NFVTime OS is based on various Openstack components to allow support vCPE lifecycle, security and manageability of the vCPE
  • VIM is integrated with all the basic components (Cinder, Glance, Nova, Neutron) – running as agents. It can be  controlled by either a remote centralized OpenStack Controller, or with a fully integrated local OpenStack Controller running as part of NFVTime-OS. In this case the vCPE can be managed via standard OpenStack REST APIs.
  • NFVTime-OS supports Backup & Restore functionality in conjunction with NFVTime-manager and Provisioning Server component allowing periodic backups and automated disaster recovery and node replacement.
  • Security and authentication, NFVTime-OS  uses central Access Control function for the OS, including Keystone that authorizes all API & CLI access, generated by human or automated systems (e.g. Orchestrator), via its local DB or by accessing external AAA systems, such as LDAP, TACACS+ or RADIUS

Control Plane Apps provide enhancements for routing and forwarding of management traffic within the vCPE, as well as monitoring and testing for all traffic.
  • BGP routing for management traffic. Management traffic is separate from data traffic by OVS, based on VLAN. IPv4 is supported for management traffic.
  • TWAMP for L3 performance testing
  • SNMP, mainly for performance monitoring – IF-MIB, RMON, and Health Monitor alarms can be retrieved via SNMP
  • Syslog – for storing any OS logs on a remote syslog server
vCPE Lifecycle Management Apps provide functions not available in OpenStack such as:

  • Security – includes platform security, SSO, remote secure upgrade and more
  • vCPE Health Monitoring agent – monitors vCPE health on different levels, and provides real time notifications / alarms to the Orchestration / Controller systems when an issue is detect. In some cases it attempts self-healing. The following monitoring levels are supported:
    • HW level – temperature, fans, PSUs, CPU/RAM/Disk usage, etc.
    • OS level – critical Linux processes & services, OVS, KVM, etc.
    • OpenStack level – OpenStack agents (Nova, Neutron, etc.)
    • Virtual Network level –  virtual network traffic statistics (per VLAN, per flow, per virtual port)
    • VM level – aliveness of VM processes, and possibly their vCPU/vRAM usage. This doesn't include monitoring of the VNF itself (e.g. vFW internal statistics). Also, the Health Monitoring agent will never attempt VM self-healing as this must be controlled by the Orchestrator or VNFM. 
  • Home Calling & Remote Upgrade agent – a real Zero Touch Provisioning functionality. It connects to the default Provisioning Server upon first boot and initiates the Home Calling procedure. 

Build for easy integration, Unified Management Layer

All management & control functions built in addition to OpenStack components can be controlled via unified CLI or through Telco Systems REST API. 

The Unified Management Layer provides the northbound REST APIs (both Telco & OpenStack) to any Orchestrator or Controller and also connects to the Provisioning Server. In any addition, NFVTime-OS provides a unified CLI built on top of the REST API layer so it can easily support any function provided by the REST APIs (Telco or OpenStack). 

In addition, OpenStack Horizon GUI can be enabled on the vCPE if it runs with a local OpenStack Controller.