The freight train of cloud computing has left the station and is on an exponential growth tear. But just like in the cartoon strips, there might be broken tracks or a chasm ahead with emerging complexity and ensuing fragility in the public cloud. In terms of cloud computing, should enterprises enjoy the fast paced ride, or think cautiously about what lies ahead?
The freight train of cloud computing has left the station and is on an exponential growth tear. But just like in the cartoon strips, there might be broken tracks or a chasm ahead with emerging complexity and ensuing fragility in the public cloud. In terms of cloud computing, should enterprises enjoy the fast paced ride, or think cautiously about what lies ahead?
Cloud computing makes sense for companies seeking to deploy IT solutions faster and more flexibly on a pay-as-you go basis. However, it’s not all upside for cloud computing as complexity in cloud environments (especially public) increase, there’s also potential for catastrophic failure within systems. And while IT and business users should expect occasional downtime, as public cloud complexity rises there’s also potential risks for much worse.
Previous columns have examined architectures and technological complexities of cloud computing environments. We’ve also examined how the moving pieces in cloud computing are often interdependent and tightly coupled where failure in one component can affect the performance of others. We have also seen how it’s wise to not assume large cloud providers offer a safer choice in terms of keeping data secure, and protected from loss.
As cloud environments inevitably experience technological advances, increased multi-tenancy, colossal system sizes, tight coupling of processes and components, and myriad IT personnel decisions (and errors) these systems will grow more risky to the point where system accidents will become commonplace.
Charles Perrow, author of “Normal Accidents”, says of such environments; “Given (these system) characteristics, multiple and unexpected interactions of failures are inevitable. This is an expression of an integral characteristic of the system, not a statement of frequency.”
Solutions then to these challenges then include adding more redundancy and buffers for components and processes and also additional data protections (i.e. backups on and offsite) to prevent temporary or worse – complete data loss. It should also be a goal to lessen the chances of failure through better training of personnel managing such systems and disaster planning with the expectation that system failure isn’t just possible, it’s very likely.
Public clouds are extremely beneficial to many organizations–allowing them to obtain compute and storage resources with just a few clicks and a credit card. However, there are certainly risks and other considerations (e.g. operations, data, security, privacy, legal) as well–and these should not be overlooked.
Perrow reminds us that great events have small beginnings. With data as the lifeblood of an organization, it’s good to enjoy all the benefits that cloud computing brings, however it’s also wise to pay attention to the little details and dependencies that could turn a small hiccup into a severe case of heartburn.
