Interval Merging Binary Tree

Abstract

The general area of the paper is methods and data structures to efficiently avoid data duplication. In telecommunication networks operation support systems (OSS) process time series of counters related to the behaviour of network elements, such as failed location updates over the last 5 min. In general we may assume time series of key-value pairs with the key encoding the ordered sequence number of the particular counter.

In certain scenarios packets are duplicated in the course of transmission from the network element to the OSS system. In other scenarios packets arrive out of order and some of them do not arrive at all. As a result KPI-s aggregated from the individual counters held by the packets will have incorrect values potentially resulting in thresholds agreed in SLA-s being falsely exceeded. The filtering of duplicated keys and the management of missing (out of order) keys should operate fast and exhibit relatively low memory footprint. For this purpose well known data constructs like hashes or binary search trees can be used, but usually they need to store all individual keys. This implies high memory footprint and slow operation, since the time complexity of a search or insert operation is proportional to the number of stored elements in most cases. We propose a special type of binary tree that overcomes both limitations within certain constraints.