In August 2014, we hit a critical milestone metric impacting how traffic is moved about the Internet. This went mostly unnoticed, however some organizations experienced outages for a brief time in relation to this event. Detailed information is available in this article on hitting an internet routing milestone. The issue generally impacted older network routing equipment, but highlights the need to keep abreast of general trends on internet usage.
For this issue, the impact was mostly felt at small ISPs with older routing equipment or at their providers if they are in turn being serviced by older equipment. The last entrant in the routing table was effectively dropped, which makes the associated servers effectively unreachable for traffic passing through the affected router. The fix for the ISP or entity managing the effected router was relatively straightforward, but appears to impact rollout and adoption of the newer IPv6 protocols for all clients of those ISPs – at least until they upgrade their routing equipment.
A point of continuing interest that bares monitoring is the mention in the article of the exhaustion of the IPv4 address space causing fragmentation of the existing address space to accelerate. This apparently will cause us to hit the next milestone that much sooner, accelerating the timing for risk of a similar outage in the future.
Considering the aging Internet infrastructure, the slow uptake on newer protocols, and the challenges associated around these milestones – we consider our recent move to cloud based infrastructure as an effective mitigation against future risk along these lines. The large cloud providers generally controls the routing infrastructure for very large IP address blocks with much less fragmentation in the routing tables, hence they statistically have less exposure to these types of issues due to the scale at which their infrastructure is managed and the presence of more up-to-date equipment where they are closer to or directly on the Internet backbone.