OpenStack NFV

(Network Functions Virtualization)

As telecommunications companies prepare for 5G, they are increasingly turning to network functions virtualization (NFV) to revolutionize the way their networks are built and delivered.

NFV allows telcos to reduce their dependency on proprietary appliances, so they can take advantage of new open source technologies on a broader scale and roll out new technologies and services in days, rather than months. NFV is crucial to modernizing telecommunications infrastructure, since it gives companies the power to make their networks as uniform as their data centers, leverage commodity hardware, and implement new automated services in an agile way.

However, the path to NFV is far from straightforward: the telecommunications industry is still ruled by regulations, complexity, and quality of service requirements. These pressures demand discipline and care from companies making the shift to NFV.

What is Network Functions Virtualization (NFV)

Network function(s) virtualization is a network architecture concept that was created by telecommunications companies to move services — like load balancing, firewalls, and intrusion prevention systems — away from dedicated hardware and into a virtualized environment. It virtualizes network nodes (typically, hardware appliances) into building blocks that can be connected to create communication services.

This concept allows operators to enjoy the many benefits of common IT virtualization, most notably eliminating custom hardware appliances for each network function. With NFV, telecommunications companies have the freedom to host VMs running on general-purpose hardware or cloud infrastructure rather than dedicated network hardware. This has the dual benefit of easing deployment as well as decommissioning: IT administrators can simply reuse the hardware elsewhere because the hardware is all the same.

This need for flexibility, coupled with the soaring costs of hardware-specific appliances, has driven the rapid adoption of NFV within telecommunications companies.

VNFs, NFVI, and Other Terms to Know

Virtualized network functions (VNFs) are related to NFV, but not the same: VNFs are implementations of network functions that can be deployed on an NFV infrastructure (NFVI). They can include one or more VMs running different software and processes.

NFV Infrastructure (NFVI) includes all hardware and software components of the environment in which VNFs are deployed, spanning from basic servers, switches, and storage devices to multiple virtualization and cloud technologies.

NFV Management and Orchestration (MANO) is where all the data is held — including all functional blocks, data repositories, and reference points — but is not part of the VM data. As the name suggests, MANO enables the management and orchestration of NFVI and VNFs.

A Virtualized Infrastructure Manager (VIM) controls and manages the NFVI. Often, this includes managing a repository and keeping an inventory of hardware/software resources, as well as organizing virtual links between VNFs.

NFVI and VIM are often inseparable. For OpenStack, NFVI is composed of the hardware and virtual services such as compute (Nova), block storage (Cinder) and network (Neutron). VIM is composed of the control plane including Horizon, the Nova controller, the Cinder controller, and the Neutron controller.

Download the Whitepaper

OpenStack Backup & Recovery Whitepaper

Learn about: the rise of OpenStack within enterprises around the world, the gaps that have been created by this shift, and TrilioVault’s cloud-native approach to filling the gaps left by legacy solutions.

Take a Deep Dive on OpenStack Backup

Learn how the shift to open source is impacting cloud data protection.

NFV Driving Modernization

OpenStack as an NFVI Platform

OpenStack has become the de-facto standard for NFV deployments within telcos. In fact, a Heavy Reading survey showed that 85.8% of telecoms regard OpenStack as a critical component to their NFV success. Some telecommunications companies that have already implemented OpenStack NFV include AT&T, Verizon, Deutsche Telekom, NTT Group, and more.

Many of these companies chose OpenStack as their NFV foundation for the same reasons it has become so popular in other industries: multi-tenancy, scalability, rich APIs, a modular architecture, and a vibrant, growing community.

OpenStack is often deployed as a virtualized infrastructure manager (VIM), but is capable of much more than that. At a minimum, telcos can use OpenStack to control and manage their compute, network, and storage resources.

However, OpenStack is preferred by telcos because it also: delivers multi-tenant management and resource sharing capabilities, provides platform resiliency with a highly available control plane, exposes APIs to enable DevOps/lifecycle management, and enables the collection of fault and performance data for physical and virtual resources.

However, in and of itself, virtualizing network appliances isn’t a fix-all solution. By itself, NFV will reduce your capital expenditures by enabling the usage of repurposed or commodity hardware, particularly if you’re deploying on bare metal (a capability introduced in the Queens release). The real savings comes via running VMs on a cloud, which allows your company to optimize the entire network stack. That’s the part that can amount to significant savings in infrastructure and maintenance costs.

NFV Enabling 5G Infrastructure

5G is much more than an upgraded replacement for 4G LTE. It’s a new approach to infrastructure that introduces modularity and flexibility to an industry that’s previously been dominated by rigid, purpose-built hardware.

And the promise of 5G is compelling. Verizon, for example, has announced plans for a wireless internet service for home use, built on the 5G network. Similarly, Ericsson recounted their lofty goals and requirements for 5G networks:

  • End-to-end latency of a few milliseconds
  • 10 years battery life
  • 10 Gbps data rate
  • 1000x increase in mobile data volumes
  • 10-100x increase in connected devices

However, the infrastructure, flexibility, and scale demanded by 5G networks adds a layer of complexity and introduces all-new requirements for technology teams at telecommunications companies around the world. How can they scale to meet the massive traffic and device growth that is expected from the introduction of 5G? Enter, NFV.

NFV can be used to enable the virtualization of networks of all kinds, including communication services like the highly-anticipated 5G networks. In 5G, NFV will enable network slicing, which allows multiple virtual networks to be created on top of a shared physical infrastructure. This enables different types of applications and services to run on shared infrastructure, creating a user experience that’s indistinguishable from a physically separated network, similar to tenants environments within OpenStack.

OpenStack and NFV are very intertwined here, too. According to the same Heavy Ready survey, more than 30% of telecoms are currently using or testing OpenStack for 5G. In fact, according to Red Hat’s VP of Telecommunications, Darrell Jordan-Smith, “5G is the reason why telcos are building NFVI, because they realize that, to connect all of those devices to their network-based infrastructure, they need to do it intelligently. They need to do it at the edge. And they need to add a high degree of flexibility and agility to their network-based infrastructure to create an innovation environment for application developers to connect all of those devices.”

The Need for VNF Data Protection

You've just built a brand new infrastructure that adheres to the NFV standard — now you need to protect it. If you are only deploying brand-new virtual functions that are designed as stateless applications, then your VNF data is centralized and is (hopefully) backed up elsewhere. However, the vast majority of companies don't have the luxury of rewriting their entire functional layer, and end up porting over at least some of their legacy applications. Those workloads are at risk if they're not being backed up.

So how do you know what needs data protection? In short, any network function that stores data needs to be backed up. This can include both operational data (device configurations, service parameters, policies) and collected data (logs, analytics, usage records).

NFV deployments need data protection for the same fundamental business reasons as other enterprise environments: to enable users to restore workloads quickly in case of data damage, data loss, or an area-wide disaster. Since NFV is complex by nature, it takes time and effort to tune VNFs correctly. Consequently, although it is possible to launch stateless VNFs from scratch, it it is considerably faster to recover VNFs from a well-known point-in-time backup.

Unprotected VNFs: What Can Go Wrong

Disaster can strike in many forms. Sometimes it may cause the loss of specific VNFs, while other times it may compromise an entire data center.

If your organization stores legacy applications or VMs within unprotected VNFs, they are subject to a number of risks both inside and outside the VNF. Administrators can unknowingly alter configurations or networking definitions, delete files, and delete VMs. Software updates and errors can cause data corruption. Automation scripts can overwrite VMs or data. The list goes on.

The point is, any number of errors and mistakes can dramatically jeopardize the integrity of your VNFs and require emergency roll back or disaster recovery measures. A data protection solution for OpenStack NFV can mitigate these risks and position you for success.

TrilioVault: Native OpenStack NFV Backup

TrilioVault is the only OpenStack-native data protection solution on the market. Check out what makes TrilioVault so unique.

Non-Disruptive Backup

Schedule application-aware snapshots to capture changed blocks of workloads via an agentless Data Mover

One-Click Restore

Recover to a point-in-time from a synthetic full image of your environment, created from incremental scheduled backups

Self-Service Management

Tenants can restore individual items, VNFs, full workloads, or entire cloud environments on-demand via the Horizon GUI

The Only Native Backup and Recovery Solution for OpenStack NFV

With multi-tenant support and seamless OpenStack integration,
TrilioVault fits into your existing NFV cloud workflows.

TrilioVault makes recovery for OpenStack NFV much easier. TrilioVault's application-aware snaps capture entire workloads to bring your environment back to present day quickly and efficiently.

TrilioVault Makes OpenStack NFV Recovery Easy

TrilioVault’s application-aware snaps capture entire workloads including the VMs, flavors of the VMs, network settings, and all the volumes attached to those VMs. With DevOps tools plus historical fixes, patches, and tweaks per machine, you can bring your environment back to present day quickly and efficiently without the need for low-level scripting or manual refactoring.

Protecting your OpenStack NFV environment enables you to recover from data damage or disaster significantly faster than by re-deploying an entire VNF from scratch. TrilioVault captures entire workloads (including applications, OS, network configuration, storage volumes, and other critical metadata).

TrilioVault enables business-critical services like NFV to be quickly configured to a previously working state, assuring an optimal quality of service. VNFs are treated the same as any workload in TrilioVault and are backed up in their entirety, including applications, VMs, networks, and data/metadata. You can:

  • Define a protection policy, backup schedule, and retention period
  • Recover an entire VNF, selected VNFCs, or individual volumes, networks, etc. to a point-in-time working copy
  • Restore an entire workload to the same or another availability zone/tenant
  • Restore multiple workloads to another cloud altogether for disaster recovery

TrilioVault for OpenStack NFV Reference Architecture

TrilioVault for OpenStack NFV Reference Architecture

Protecting your OpenStack NFV environment enables you to recover from data damage or disaster significantly faster than by re-deploying an entire VNF from scratch. TrilioVault captures entire workloads (including applications, OS, network configuration, storage volumes, and other critical metadata).

Covers Critical OpenStack NFV Use Cases:

  • Captures point-in-time snapshots of VNFs, so users can restore complete environments including VM flavors and network topology/configurations
  • Supports deployment via Red Hat Director, Juju, and other lifecycle management tools, so administrators can deploy the two solutions together using a single tool
  • Leverages incremental forever snapshots that include changed block tracking
  • Supports Ceph S3 cloud storage, and is certified for Red Hat Ceph
  • Covers both backup/recovery and disaster recovery scenarios

Enterprise-Ready

  • Highly available
  • Tenant-driven & self-service
  • Administrative oversight of backup scheduling & policies
  • Role-based access control
  • Keystone & trustee role authorization
  • Chargeback for tenant usage & consumption
  • Software-only: no additional hardware required
  • Compatible with all leading OpenStack distributions

Articles Related to OpenStack NFV Backup