Scale and Other Requirements in the Data Center and Campus
In a recent blog, we discussed connectivity options for enterprise data centers building hybrid clouds. One of the options was to connect cloud providers directly over the Internet, which has the advantage of being able to use an existing Internet connection–it’s the easiest option and you can connect to any number of cloud service providers.
Following that article, we wrote about modern data center scale requirements, covering interface density and other requirements for data center solutions, whether housed on enterprise premises or in colocation facilities. Here, we’ll focus on Internet connectivity and some of the requirements, including scale.
Internet connectivity from enterprise data centers is growing in importance for multiple reasons. These include the Internet of Things (IoT) or the “bring your own device (BYOD)” trends, further complicated by the desire for content and mobility. This all leads to higher routing table sizes, especially if routes need to be disaggregated for more specific reachability and better performance.
Requirements for Border Routing
Internet routing may also be referred to as edge or border routing. The edge, or border, is the place in the network that aggregates the customer, service, and Internet connections in and out of the data center or campus. Figure 1 shows Internet Border Routers connecting to Internet Service Providers (ISP). The underlying network they are connecting could be an IP or an Ethernet fabric, a core-aggregation data center, or a campus network.
Figure 1: Connecting to Internet Service Providers from a Data Center or Campus Fabric
High availability and redundancy are important considerations here, since traffic loss at this location could take important services or users offline.
But what are the requirements for edge routing from an enterprise data center? All of the major IPv4 and IPv6 protocols (interior as well as BGP) must be supported, as well as MPLS VPNs and multicast. If multitenancy is a consideration, it is important to support logical domains through constructs such as VPLS or BGP-EVPN with VXLAN.
Finally, visibility and automation are key requirements here as in all network locations. This allows administrators to verify and prove strict service level agreements (SLA) with their users. Relevant data collection should include interface states and statistics, flow information, and the ability to examine and take remedial actions based on individual packets.
Internet Routing Table Growth
For reasons discussed above, the internet routing table is growing. Figure 2 illustrates growth since 2010, through the present and on to 2030.
Figure 2: Historical and Future Growth of Internet Routes (Source: https://www.potaroo.net/)
As you can see, Internet routing table growth could top 1M routes by around 2020 and scale to 1.5M by 2025. Accordingly, it would be prudent for a routing solution to support that scale today.
Extreme border routing solutions include the SLX 9540 and SLX 9850. SLX 9540 is a fixed form factor platform that enables flexible edge connectivity for data center interconnect, WAN edge, IXP, and colocation data center deployments. SLX 9850 is a chassis-based platform that delivers industry-leading 100/40/10GbE port density and scale. Both the SLX 9450 and SLX 9850 have embedded network visibility through the SLX Insight Architecture.
Figure 3: SLX 9540 and SLX 9850 for Scalable Border Routing
These platforms support all of the requirements detailed above for Internet routing. In particular, they support a table size of 1.5 million IP routes today, ensuring headroom for the next 5-7 years. This scale is made possible through our new technology called Extreme OptiScale™ for Internet Routing that optimizes programmable hardware and software capabilities to accelerate innovation and deliver investment protection.
Specifically designed to address the evolving technical and business needs of Internet border routing, Extreme OptiScale™ for Internet Routing delivers the most efficient solution today for Internet-scale border routing, and this functionality is extensible to other constructs such as access lists and VPNs.
This post was originally published by Product Marketing Director Alan Sardella.