My previous post pondered whether unlimited data could limit data silos, which was inspired by an extended disruption in the internet service provided by my local cable company. This provided me the opportunity to see just how unlimited the unlimited data plan on my smartphone really was. Ironically, shortly after my cable internet service was restored, I received the following email alert from them about my data usage:
Until last year, my cable company had an unlimited data plan by default since the only pricing differences for internet packages were based on data streaming speed. Then they decided that 250 GB a month was the default limit, after which I get additional data in 50 GB blocks billed at $10 per block. As the above chart shows, although I was well below that limit for the three previous billing cycles, my extra data streaming of movies and television shows on Netflix during the holiday break was pushing me close to my data limit.
Could unlimited data lead to unlimited data silos?
My previous post posited that the enterprise data management equivalent of data streaming services would limit data silos by preventing users from essentially downloading copies of enterprise data, most of which they had a very short-term use for anyway. However, while streaming limits the long-term storage and maintenance of redundant copies of data, an unlimited enterprise data plan would probably be unrealistic.
As Alan D. Duncan commented on my previous post, the underlying data management and infrastructure capabilities needed to enable enterprise data streaming are significant. Therefore, limits would probably have to be enforced on the amount of data users could stream, as well as perhaps limiting their data streaming speed. And the frustrations caused by these limits could ironically encourage the creation of even more data silos.
To infinity and beyond
“If I were in charge of naming,” Mark Horseman commented via Twitter, “I’d call it infinite data. Data potentially grows faster than you could stream it.” To which I replied that you only need to stream as much as a user can consume, just like streaming a two-hour movie only buffers a rolling 15 minutes of data at a time.
However, while big data is as infinite as a never-ending movie, our time remains finite. Complicating the plot further, often at the big data multiplex crisscrossing data streams are simultaneously broadcasting – how many different movies can you watch at the same time?
“I suppose the half-life of data would be a consideration too,” Horseman added. “Does it take so long to stream that decisions are no longer valid?” Another excellent point, but I argue it’s the human decision speed limit that’s the real bottleneck, which is why data streaming often relies heavily on algorithm-driven decisions.
What stream you?
What does your stream of consciousness have to say about the limits of data streaming?
