A case for redundant arrays of inexpensive disks (RAID)
SIGMOD '88 Proceedings of the 1988 ACM SIGMOD international conference on Management of data
Performability Analysis Using Semi-Markov Reward Processes
IEEE Transactions on Computers
Power laws and the AS-level internet topology
IEEE/ACM Transactions on Networking (TON)
IEEE/ACM Transactions on Networking (TON)
Hi-index | 0.00 |
Random Graph (RG) model is based on the notion that statistically all nodes have identical connectivity pattern. Given n nodes and k randomly distributed edges (k≪ n-1)/2) leading to a bell shaped distribution with a mean=k/n. Actual measurements performed on large networks showed that connectivity behavior follows a power law distribution instead. This distribution causes large networks to exhibit the behavior of preferential attachment to nodes with high degree. Furthermore, this behavior of preferential attachment is exhibited by networks of varying sizes, from thousands of nodes to billions of nodes, leading to a scale free graph (SFG) model. This paper analyzes the effects of multiple failures on networks modeled as RGs and SFGs by measuring network fragmentation and diameter as a function of the number of failures. We also propose a new routing technique, called the Redundant Array of Inexpensive Links (RAIL) that makes connectivity inherently tolerant to multiple failures.