Consideration as things move to the “Cloud”

This morning’s mail brought a news e-mail with an article by Mike Fratto Bandwidth Management Coming To You - Analytics - InformationWeek. This struck a chord with me as I have been thinking about (and participating on the edges of) cloud computing and services lately.

Putting software and services in the cloud is compelling. There is less dependence on the client computer and the maintenance efforts associated with that moving to more centralized support without having to worry about touching the desktop. Additionally, taking care of disaster recovery, relocation of the desktop (like on a trip or a vacation or changing operating systems), and backup can be done centrally.

The downsides include privacy concerns (although they are with us with desktop centric software as well), the ability to “work” when disconnected from the net, and large impacts when an outage does occur.

Mike Fratto’s article (that motivated this post) raises another issue – upcoming network congestion. The article focuses on Cox Communications’ bandwidth management announcements as well as the thought that all networks are going to have this problem.

Thinking over my last couple days:

• Noted my university detailing it’s move from network pipes (more than one) that are limited to 300Mb/s to those supporting 1Gb/s (although our usage is not filling the new pipe yet!)
• Watched a “watch instantly” NetFlix movie over the net
• Moved 1/2 my e-mail to the Google cloud
• Heard that IEEE as well as many other organizations has adopted VOIP
• Patched my iPHone (260MB update)
• Seen (additional) presentations where SPAM is 90% of the mail traffic (and seemingly unstopped in transit)
• Set up collaborative space for an ad hoc committee on the cloud
• Noted companies are continuing to pull in their efforts as the economy continues in a recession
• Noted the many discussions about outsourcing e-mail, document processing, etc. to cloud applications

and noting that I am probably at least a bit in front of the general adoption curve, the idea that we may move into a period of bandwidth limitations seems very probable.

Are bandwidth limitations going to have a significant performance impact on cloud applications? It is not clear. The cloud application providers are paying for bandwidth as they use it (in large chunks) but it is less clear what the (home) user is really paying for. The commercials are all related to peak not available and not realized. The non-net neutrality folks would let the communications vendors prioritize traffic based on financial deals that will create zones of good performance for some applications for some people. The Cox Communications approach is more of the “net neutrality” approach trying to prioritize based on protocol and its (expected) impact on the user experience.

Updated 1/30/2009 in evening to add “changing operating systems” to the flexibility given by using the cloud.