Ryan Smith is a network engineer for Cervalis LLC, a premier provider of IT infrastructure and managed services.
In most networks the routers that are used to peer with upstream providers are mostly considered a set-it-and-forget-it type of device. Substantial work and testing is done initially to establish peering with the ISP and then, for the most part, administrators are free to step back and let the routers and routing protocols take over. What many don’t realize is that these routers can easily turn into time bombs ready to blow at any minute.
Growth of the Internet route exchange
Border gateway protocol (BGP) is, by and large, the most predominant method of exchanging routes with a provider. Through this protocol organizations not only are able to accept and transmit routes, but also to modify and manipulate the routes they receive. This peering process is one of the fundamental building blocks of the modern day Internet.
With the rapidly approaching complete depletion of IPv4 addressing space, large ISPs, ARIN and other RIRs are constantly working together to reclaim and repurpose IPv4 resources and to segment larger, previously aggregated blocks to serve more end companies. Because of this, we are seeing the level of Internet route exchange continue to grow over time instead of seeing it stabilize. This, together with the recent surge in acceptance of IPv6, means that organizations that handle their own route exchanges need to be cautious and may need to reexamine their environments.
Back in the earlier part of 2010, the size of the global routing table was under 300,000 IPv4 routes. While large, this table was much more manageable as compared to today’s 500,000+ IPv4 routes. Combined with the sharp increase in IPv6 routes - which today number nearly 19,000 - this means that the router memory space consumed by these tables has almost doubled. With the recent scramble to reallocate IPv4 resources, an IPv4 route that previously was aggregated into a single /19 subnet representing one line in the table may soon be segmented into 32 different /24’s taking up 32 lines in the table. Coupled with the ongoing growth of IPv6, this will cause the table sizes and memory requirement to continue to balloon.
Reaching the point of critical mass
Network administrators and engineers must pay very special attention to their edge routers and peering relationships with upstream providers. Accepting this many routes from a provider may have been fine initially, but if left uninhibited, this growth may quickly consume all the available memory on a router and lead to major traffic forwarding and router stability issues.
Compounding this problem is the fact that this route growth happens completely unexpectedly. In a period of just a few hours, without warning route tables can grow by thousands of routes. Hitting the point of critical mass is something no organization wants to experience.
Defining the needs of your organization
Numerous techniques and methods can be applied both at the network edge itself and within the provider’s network. Administrators should define what their organization’s needs are when arranging a peering relationship with an upstream provider. In most instances a full routing table isn’t required and a summarized table can more than suffice. Using summarized tables organizations can exponentially decrease the table size and system hardware requirements, although they might miss out on some more advanced traffic management methods. When full routing tables are required, administrators should take a close look at the resources available on their routers and determine whether or not an upgrade is needed.
Certain managed services providers are in a unique position to help customers permanently eliminate this time bomb issue from their network as they regularly deal with aggregating multiple gigabits of traffic to upstream and thus maintain substantial edge networks to route and manage the traffic.
These routing platforms are purpose-built with a service provider mindset and for which leading edge hardware and technology have been deployed. Because of this they are able to maintain full routing tables from numerous providers simultaneously and can easily deal with the continual growth of the global table. Customers of these managed services providers are shielded by this routing table growth and can also leverage additional routing capabilities and options through the MSP that can’t be achieved by working directly with an Internet provider. Routing techniques and methodologies that were previously unobtainable on smaller routing platforms can easily be attained on an MSP’s platform.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.