Consider these statistics from the Cisco Visual Networking Index Complete Traffic Forecast (2016-2021):
- Global IP traffic will reach 3.3ZB (zettabytes) per year by 2021. That’s 278EB (exabytes) per month. An exabyte is 1 million terabytes and a zettabyte is 1,000 exabytes.
- IP traffic will see a compound annual growth rate of 24 percent through 2021, to reach 3.2 times 2016 levels.
- Traffic during the busiest hours of the day will grow even faster, increasing by a factor of 4.6 from 2016 to 2021.
Growing IP traffic volumes are forcing cloud and content providers to rethink their wide-area network (WAN) architectures. Traditionally, organizations would connect to an ISP’s point of presence (POP), which provides access to upstream Tier 2 and Tier 1 networks. Large ISPs typically have POPs in major cities, all of which are interconnected by the ISP’s dedicated network backbone.
As network traffic continues to increase, this design creates chokepoints due to oversubscription. In addition, the distance between the end-user and the Internet backbone causes latency that impacts application performance and content delivery.
The new approach is for organizations to build their own POPs to bring applications and data closer to users and create a more resilient distributed network. A POP is a micro data center outfitted with servers, routers, switches and other networking equipment that enables connectivity with upstream providers.
Although POPs are small, they still require careful planning and attention to detail. The first step is to select a physical data center that offers the right blend of performance, security, certification and, of course, cost. Because the primary function of the POP is connectivity, the ideal data center will offer multiple data transit options and direct connection to an Internet exchange point.
The next step is to design, configure and implement the POP equipment. Again, the POP should be architected with high-performance connectivity in mind. It’s also important to build in security, particularly to defend against distributed denial of service (DDoS) attacks. The equipment should be preconfigured and tested then shipped to the data center with rack elevations and cabling schematics. There the equipment is “racked and stacked” and the POP brought online.
It sounds simple enough, but few organizations imagined they would need to get into the ISP business. Furthermore, growing cloud and content providers will typically need to build out multiple POPs across a wide geography. This requires a strategy for replicating the POP architecture quickly and efficiently.
Rahi Systems offers a number of solutions and services that facilitate POP implementation. Our FlexIT POD and OIB (Office-in-Box) Rack incorporate cooling, smart power distribution, backup power, network services and physical security in a self-contained unit. These solutions reduce the complexity of the data center infrastructure buildout and provide a uniform design that enables rapid deployment of IT equipment. The FlexIT POD also features a unique in-row cooling system that can support up to 100kW of load.
The Rahi team has extensive experience in the planning, deployment and support of POPs for some of the world’s largest cloud and content providers. In our next post, we’ll discuss how our professional and managed services can help ensure that your POP is built out properly and kept operating efficiently.