Reduce CapEx While Maximizing Performance with NFV

Cloud technologies have transformed not only the provisioning and delivery of computing services, but also IT business and financial models. Enterprises have delegated management of an ever-increasing amount of their IT infrastructure to cloud service providers, due to the realization of significant capital expenditure (CapEx) savings through the economies of scale delivered by the cloud.

In many instances, the cloud has reduced, if not eliminated, the time and cost of procuring, installing and managing physical on-premises IT assets—servers, storage, communications and rack space. As data and processing demands increase or services need to be provided in new locations, the cloud has enabled these resources to be rapidly and economically deployed.

Virtualization applied to network functionality

Virtualization—the concept of separating the software functionality from the specific hardware on which it runs—has played a major role in the adoption of cloud services. Virtualization has helped drive down IT infrastructure costs. Virtualization is now being applied to other areas of IT infrastructure—specifically, to network hardware. As a result, enterprises needing to extend their network reach to new markets or provide new services can now achieve these goals far more efficiently.

Network functions virtualization brings CapEx advantages

The capabilities of some of the most common yet critical network hardware devices used in cloud environments are being virtualized. The functionality of traditional network functions such as routers, firewalls and load balancers are now available virtually through network functions virtualization (NFV) software deployed on commodity servers. By eliminating the need for purpose-built network hardware, virtualization brings considerable CapEx and operational advantages. Contrast the deployment cycle of physical network hardware with an on-demand provisioning process for virtual network devices virtualization software.

The process of deploying physical network hardware and connecting it to your cloud services  typically involves the following steps:

  • Evaluate and select network hardware vendor and device (usually takes days to weeks)
  • Submit a purchase order via your organization’s procurement process (usually takes weeks)
  • Install the network device in the data center (usually takes days)
  • Configure the network device (usually takes a day)
  • Order ports (usually takes days)
  • Establish connections to your cloud environment (usually takes minutes)

The entire process could take several weeks or even months and requires multiple lines of business to complete the task. The process becomes all the more complex when multiple devices must be procured for deployment in different locations across a widely distributed network infrastructure.

In contrast, provisioning and deploying a virtual environment is dramatically easier and faster:

  • Access the provider’s portal (usually takes seconds)
  • Select and purchase the virtual network function (usually takes seconds)
  • Deploy the virtual network function on the provider’s hardware (usually takes minutes)
  • Connect to your cloud environment(s) (usually takes minutes)

The virtualized approach provides rapid deployment and replaces the up-front CapEx associated with hardware acquisition with OpEx software licensing.

The challenge of connecting between public clouds

It’s rare that a single public cloud service provider offers all of the functionality, geographic coverage and application and data support services an enterprise requires. According to the RightScale 2018 State of the Cloud Report, on average an enterprise uses 4.8 different cloud environments. Applications in one cloud environment often need to access resources in another cloud environment.

One means of establishing a connection between public cloud environments is via the public internet. While this is a low-cost approach, it generally doesn’t meet performance requirements. The public internet does not guarantee a reliable, consistent connection between cloud environments. As a result, latency is less than optimal and leads to variable processing performance. A public internet connection also presents a less-than-secure target for hackers.

A more secure approach is through a private connection between public clouds using a physical network router. This provides a reliable dedicated connection and ultra-low latency for optimum performance. However, as mentioned above, the network hardware procurement process precludes rapid implementation. In the complex multi-cloud environments that characterize most enterprises, particularly those with global reach, that approach is untenable.

Virtual Networking for cost-effective connectivity between public clouds

The best approach for establishing connections between public clouds uses a virtual networking approach. Rapid deployment and lower CapEx in comparison to physical network hardware devices are key advantages. In an actual NFV example, a Cisco Cloud Services Router 1000V deployed in one of our global data centers connects applications running in AWS with databases managed on Oracle’s Cloud Infrastructure. Deployment time was minutes, not weeks. More importantly, this architecture delivered network round-trip latency of sub-2 milliseconds between the Oracle Cloud Infrastructure and AWS.

For enterprises that have widely-distributed network infrastructure and contract with multiple cloud providers, the efficiency of deploying network functions without the need to procure and provision hardware is very appealing—from a financial as well as operational perspective.

NFV availability through Platform Equinix

The future of Platform Equinix will aim to provide an environment to deploy virtual network devices and connect them via software-defined interconnection. This global infrastructure will enable enterprises to securely route between cloud environments and extend network reach to new markets worldwide. By taking advantage of the ease of deploying virtual network devices, enterprises can reduce CapEx and rapidly provision network functions across distributed enterprise networks to maximize performance.

To learn more about how your enterprise can benefit from Platform Equinix and network functions virtualization, visit https://eqix.it/nfv.

 

Print Friendly


Related Content