NFV vs. SDN for OpenStack
Networking provides the vital connections necessary to run a data center. Being able to virtualize all the network functions from the command center provides clarity, improved agility, and overall efficiency to the business — all crucial elements of the next generation data center. Network Function Virtualization (NFV) is rapidly evolving how networks function — whether it be deploying or maintaining — through a number of connected parts, from routers to firewalls. Since resources can be provided dynamically, NFV lends itself well to a cloud deployment and is a burgeoning use case for OpenStack.
NFV allows physical network appliances to become software running within virtual machines or containers. NFV is similar to and derived from Software Defined Networking (SDN) but focuses more on Physical to Virtual (P2V) transformation of network functionality. NFV decouples the software from the hardware making the functions of the network independent from the hardware it resides on. NFV can aggregate usage across locations, creating immediate efficiency and costs across the organization.
Traditionally, network functions ran on specifically assigned hardware. Upgrades, additions, and changes required an operator to visit your site and physically install a new system. With NFV, the process of deploying a network now takes moments instead of weeks. The entire network, even across different locations, can be virtualized in a similar way to how a machine can be virtualized. Network functions, such as load balancers, proxies, and firewalls, are converted from their specialized hardware appliances into virtual machines (VMs) running in an orchestrated cloud. Instead of deploying a large function, such as a firewall, lots of little VMs are scaled out horizontally across the infrastructure to replace the giant firewall appliance. NFV can work with SDN to provide a number of benefits, including: economies of scale, vendor choice, faster innovation, quicker refresh, and more flexibility.
NFV and SDN: better together
Networks have been a complex beast to manage. In order to simplify management, Software Defined Networking (SDN) was built as a means to automatically manage a large number of network services and devices in a multi-tenant environment. SDN decouples a system’s control from its forwarding plane, while utilizing commodity hardware and third-party software. From SDN, the need to virtualize the network functions was identified. And so, although NFV was born from SDN, these network functions working in tandem deliver the best-in-class control and flexibility of a networked system.
Service providers, namely telecommunication companies, have been implementing NFV and SDN for a while but with limited scope. As the functions move to more flexible, API-driven software, new opportunities for innovation are emerging quickly. With some large telcos now designating up to $1 billion spend in this category, those not pursuing NFV are getting left behind.
Utilizing a NFV approach
In order for NFV to work at scale, it must be: resilient, orchestrated, elastic, and, of course, cheaper for the business. OpenStack developers have created Tacker to deploy NFV, much as Cinder deploys block storage. Tacker controls deployment of virtual network functions within the cloud.
Here is a deeper dive into two different instances when utilizing an NFV approach is appropriate for your OpenStack solution:
- When the economics make sense. NFV is really powerful for organizations with large networks that can take advantage of economies of scale. Since NFV virtualizes a lot of P2V commodity hardware, purchasing larger quantities of generic hardware can reduce both CAPEX and OPEX. Plus, assets can be recycled to build a larger cloud that deploys next generation platforms. Vendor choices also expand without specialty hardware lock-in. Small companies, too, may find it difficult and expensive to build their own hardware appliances. Small savvy companies can write disruptive NFV software to enable both small and large organizations.
- To speed up innovation and increase flexibility With no hardware to build, applications, network devices, and new services can start running sooner. The entire network becomes more flexible, scalable, and elastic. Especially as specialized hardware may become obsolete more quickly, refreshes are much more seamless, with no new specialty hardware to wait for — the commodity hardware is ready to be used as workloads shift among current resources.
Why SolidFire for NFV on OpenStack
NFV is great for optimizing your data storage, but utilizing SolidFire with NFV really enables complete orchestration across your IT environment — we believe there is no better way to do it. We’d love to share all the amazing things our customers are developing in elastic clouds with NFV and SolidFire, but they made us promise to keep our lips sealed. (But we promise they are AMAZING feats in storage optimization!) SolidFire’s geometry, resiliency, elasticity, and integration are ready to improve your company’s time to value with virtualized network functions.
First, our platform is already optimized for a virtualized workload, with lots of tiny VMs running simultaneously through its QoS. We were made for this. Second, SolidFire is resilient — its Quality of Service and self-healing allow functions to run without concern of the network going down. SolidFire has your back so you can keep forging forward with the challenges ahead of you, not behind you. Third, SolidFire’s incredible elasticity allows you to scale to your needs, whether that is up or down. Its scale-out architecture allows you to meet demands in a simple deployment manner, whenever you are ready. Finally, SolidFire’s fully API-driven capabilities make for an incredibly strong cloud platform since they were built with OpenStack solutions and containers in mind. We want to make sure your organization keeps propelling forward with the speed of development and remains ahead of the curve.