There is a very clear and marked evolution happening within the modern data center. We discuss this often, but some of these biggest shifts are actually being driven by the end user and the organization. The industry is moving toward a much more agile data center and business ecosystem. We’re seeing more data being passed through the data center, more users coming in, and a lot more focus on resource efficiency.
As it stands, global cloud traffic crossed the zettabyte threshold in 2014, and by 2019, more than 86 percent of all data center traffic will be based in the cloud, according to the latest Cisco Cloud Index report.
Significant promoters of cloud traffic growth include the rapid adoption of and migration to cloud architectures and the ability of cloud data centers to handle significantly higher traffic loads. Cloud data centers support increased virtualization, standardization, and automation. These factors lead to better performance as well as higher capacity and throughput.
As the report discusses, scalability and allocation of resources are the major advantages of virtualization and cloud computing. Administrators can bring up virtual machines and servers quickly without having the overhead of ordering or provisioning new hardware. Hardware resources can be reassigned quickly and extra processing power can be consumed by other services for maximum efficiency. By taking advantage of all the available processing power and untethering the hardware from a single server model, cost efficiencies are being realized in both private and public clouds.
Cloud computing aside, let’s dive a bit deeper into the conversation around virtualization. Specifically, network virtualization technologies. Already, managers are looking at next-generation solutions which will change the way cloud and data center resources are controlled. For example, the latest AFCOM State of the Data Center report indicates that between now and 2016, 83 percent of survey respondents said that they’ll be implementing, or have already deployed, software-defined networking or some kind of network function virtualization. Furthermore, 44 percent have deployed or will be deploying OpenStack over the course of next year. Finally, even though it’s a new technology platforms like Docker are already seeing 14 percent adoption.
Did you see that stat? 83 percent are already in some way deploying or looking at next-generation networking technologies. So, with that in mind, why are we seeing such a huge jump in interest and adoption? And, what’s the real difference between SDN and NFV to begin with?
- Understanding SDN: With SDN, at a very high level, administrators are able to control and manage entire network through the abstraction of higher-level functionality. Now, let’s dive a bit deeper. This is all accomplished by abstracting the layer which manages how traffic is distributed and where it’s being sent. This is the control plane. The underlying system helps control traffic destination. This is the data plane. To make SDN work there has to be some kind of communication between both the control and data plane, even though management is abstracted. It may sound complicated, but it really isn’t. The concept with SDN is to create a dynamic and highly programmable network infrastructure which is capable of controlling underlying infrastructure components while still being abstracted from applications and network services. This allows for better programmability across the all networking layers, better agility, central management, and an open-standards architecture. This means that SDN can drastically simplify network design by allowing administrators to aggregate physical resources, point them to an abstracted management layer (SDN) and create intelligent, programmatically configured, controls around the entire network. This means you can present network resources (via SDN) to applications and other resources. To the administrator they have visibility into the entire network flow architecture. To the applications or resources using that network resource, they simply see a logical switch. SDN’s abstraction concept fundamentally simplifies some of today’s most complicated and fragmented networking ecosystems. This is why we’re seeing so much adoption in the data center space. Organizations use SDN to deal with complexity, better policy control, improved scalability, and to remove vendor dependencies. Most of all, SDN helps with new concepts around IoT, cloud integration and cloud services, controlling vast amounts of data (big data), and even improving IT consumerization and mobility.
- Understanding NFV: Although there is a direct relationship between SDN and NFV, they’re not really dependent on each other. Network function virtualization is similar to traditional server virtualization mechanisms but clearly focuses on networking services. Within NFV, there are virtualized network functions (VNFs). It’s a confusing acronym, but an important one. VNFs are implementations of network functionality which is deployed on top of an NFV infrastructure. With that, we have orchestration capabilities, data repositories, automation layers, and management environments. All of this sits on the NFV platform. So, what are some specific examples around NFV? This could be a virtual appliance that’s only responsible for load-balancing workloads. Or, you could have a virtualized firewall scanning a specific network segment. Similarly, you can have virtual networking services like IPS, IDS, and malware engines. Finally, you can have a distributed NFV architecture using virtual WANOP accelerators as tools for network and service controls.
Catch all that? With all of this in mind one of the biggest questions still revolves around deploying SDN and NVF and understanding the use-cases. First of all, you don’t have to have both to accomplish a use-case. As mentioned earlier, these are not dependent technologies. You could very well have just an NFV platform operating a piece of your environment or just SDN.
For example, if you have a complex and fragmented network ecosystem which spans multiple data centers, it might make sense for you to abstract that control layer and involve SDN. From there, you can control network functionality, traffic distribution, and even network automation.
When it comes to NFV, let’s assume that you already have a homogenous networking environment. But, you need to control and monitor specific services within your data center. Here, you can deploy a virtual appliance which acts as a powerful load-balancer keeping an eye on workloads in your data center, the cloud, and in between. Similarly, you deploy a virtual security service which monitors traffic hitting a specific application. These are all examples of NFV deployments where you’re utilizing virtualization technologies to reduce cost and physical infrastructure complexity.
Now that you have a clearer picture, know that your organization may very well have use-cases for one or the other technology. In some cases, both might be a good fit. The key, however, is understanding the differences and knowing how SDN and NFV can positively impact your data center and your business.