Trade-offs between Consolidation and Dispersion
This equation brings us to what might be called the dispersion dilemma or perhaps the consolidation challenge. If we disperse our resources more and more, we improve response times by reducing network latency and network transport costs, but we reduce the benefit from parallelization and from statistical multiplexing. However, by consolidating, we improve our ability to parallel process. We also gain statistical multiplexing benefits. These benefits, interestingly, have an inverse square root effect, the same as for dispersion but for different reasons.
Unfortunately, we can’t get both sets of benefits at once, so dispersion versus consolidation is one of life’s trade-offs. However, customers of the cloud can leverage investments made by cloud providers to select an optimum balance, and, generally, they need not pay any more for the privilege.
Certain factors give consolidation an edge and others favor dispersion. Favoring consolidation, there are the statistics of scale, facility economies of scale, maximum manageable workloads given application coupling, and storage costs. However, user experience enhancements due to latency and response time reduction, business continuity, country compliance, customer (server hugger) comfort, favorable tax treatment, and network costs favor dispersion.
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access