BGP Churn Evolution: a Perspective from the Core

Abstract

Abstract-The scalability limitations of BGP have been a major concern lately. An important aspect of this issue is the rate of routing updates (churn) that BGP routers must process. This paper presents an analysis of the evolution of churn in four networks at the backbone of the Internet over a period of seven years and eight months, using BGP update traces from the RouteViews project. The churn rate varies widely over time and between networks. Instead of descriptive "black-box" statistical analysis, we take an exploratory data analysis approach attempting to understand the reasons behind major observed characteristics of the churn time series. We find that duplicate announcements is a major churn contributor, responsible for most large spikes. Remaining spikes are mostly caused by routing incidents that affect a large number of prefixes simultaneously. More long-term intense periods of churn, on the other hand, are caused by misconfigurations or other special events at or close to the monitored AS. After filtering pathologies and effects that are not related to the long-term evolution of churn, we analyze the remaining "baseline" churn and find that it is increasing at a rate that is similar to the growth of the number of Autonomous Systems.

Measuring the deployment of IPv6

“…Other monitors that peer with the Oregon-IX collector show similar behavior. We hypothesize that the anomalies exhibited by the AT&T and NTT monitors are caused by non-stationary periods, which we confirm by filtering out noise from the AT&T time series as described in [26] and plotted in Figure 11 (top panel) as AT&T*. Similar to the other monitors, this filtered AT&T* time series exhibits a stable ratio of 5 updates per origin AS.…”

“…The latter can be a more serious concern, because failing to process updates in a timely manner can trigger a wide-scale instability and result in traffic blackholing. Some of these concerns were put to rest by observations that churn in the IPv4 topology grows slowly [26,27], and at the same rate as the underlying topology. More recently, however, Huston [28] compared IPv4 and IPv6 BGP update time series and concluded that while IPv4 churn has grown slowly (linear), IPv6 churn has been increasing exponentially.…”

“…As stated earlier, this is likely due to denser interconnection and enforcement of business relationships in the IPv4 topology. We have studied this effect in our previous work, where we showed that correlation between IPv4 churn time series doubles after filtering out updates triggered by routing events that affect only a limited part of the Internet [26].…”

DIMR: Disjoint Interdomain Multipath Routing

“…The number of churns is the number of all messages triggered by a routing event. It is the cost of a routing protocol and also measures the scalability and instability of the protocol [18]. We observe that the difference between the PDAR and DIMR is small in terms of the number of churns, which means the control message overhead of DIMR is comparable to PDAR.…”

Route flap damping with assured reachability

“…As a result, even though persistent routing instabilities are observed repeatedly [13,8], the operational community is deeply concerned with RFD and suggested to turn it off [21]. Compounded with the fact that the MRAI timer is also being turned off at various places, the global routing system today faces a potential danger of melting down by excessive amounts of routing updates.…”