 
Summary: 687
I/OComplexity of Graph Algorithms
Kameshwar ikfunagala* Abhiram Ranadet
Department of Computer Science and Engineering
Indian Institute of Technology Bombay
Mumbai 400076 India.
Abstract
We show lower bounds of Q($ Sort(V)) for the I/O
complexity of graph theoretic problems like connected com
ponents, biconnected components, and minimum spanning
trees, where E and V are the number of edges and vertices in
the input graph, respectively. We also present a determin
istic O($ Sort(V) ' max(l,loglog 9)) algorithm for the
problem of graph connectivity, where B and D denote re
spectively the block size and number of disks. Our algorithm
includes a breadth first search; this maybe of independent
interest.
1 Introduction
Data sets of many modern applications are too large
to fit into main memory, and must reside on disk. To
