People charged with bioinformatics infrastructure development should include shared grid computing in this question of best compute resources today. While this may not be an option in Brazil, it is in the USA, UK, Europe, and Asia-Pacific, where shared grid resources for life science computing are now common, and in many cases begging for more users. The idea of buying and building your own cluster or SMP system is tempting to many of us, as it means hard equipment under our control. It also means continuing chores in maintenance and upgrades and administering those who want to use it, costs not always obvious. The agencies supporting shared grid infrastructure see the benefits of these common resources, and more of us in labs should spend some time investigating grid use as a replacement for the in-lab cluster computer systems. For many bioinformatics applications, having several of us work together to build software we all can use on grids will cut much of our duplicated efforts. For those of you who analyze genomes, please consider contributing to this Generic Model Organism Database project for shared grid computing of genome analyses (Blast, gene finding, phylogenetics, etc.) http://www.gmod.org/Genome_grid with your comments, suggestions and codes. - Don Gilbert -- d.gilbert--bioinformatics--indiana-u--bloomington-in-47405 -- gilbertd at indiana.edu -- http://marmot.bio.indiana.edu/