The Social Media Clarity podcast has just started with an initial 15 minute show devoted to analysis and advice on social media platforms and product design (http://bit.ly/12Edwvx), moderated by Randy Farmer (@frandallfarmer) and with Bryce Glass (@bryceglass) and myself. We plan to talk about a range of topics related to social media, online community, reputation systems, incentive design, analytics, visualization, and collective participation. I hope you will give a listen to the first show, Randy did a great job getting a quality recording and editing a polished package.
This episode we talk about changing rules for access to the data we all put into the cloud. The discussion is related to this post “Over the edge: Twitter API 1.1 makes “Follows” edges hard to get” – documenting the impact of changes in access to data through social media APIs. The implication from the discussion is that business may need to start building their own social network data sets since they cannot rely on cloud platforms to guarantee their access to their own data.
We plan to produce more episodes, please let us us know what topics you’d like to hear discussed.
I had an interesting discussion with Deborah Grove about the ways social media impacts the costs of computing and the efforts to “green” IT by making servers and networks more power efficient. Social media is a growing segment of all the data stored and processed on earth. But social media has different properties than previous classes of data like transactions and documents. Some social media is more resource intensive than others (a YouTube video consumes more energy to store, transmit and view than a Twitter tweet) but even light weight social media can cause heavier media types to be suddenly accessed in a high spike of demand, for example when a brief tweet points to a video that then drawn thousands of viewers.
More efficient IT is clearly important as computing consumes a growing slice of our power consumption. Social media is both a new source of demand for compute resources (and by extension, power) and a possible method for gaining new efficiency as server farms are better tuned and maintained. I would like to see major resource consuming services like YouTube and Google surface the costs of computing more. While google, for example, reports how many fractions of a second it takes for them to return your search results, it would be useful to see how many shot glasses of diesel fuel were needed to power the servers that generated those results. PCs and server farms do no have tailpipes spewing emissions but they are not carbon neutral: lots of coal, oil, gas, and uranium (and not nearly as much wind, hydo, and solar) are needed to keep the net alive. Making the energy costs of computing more visible would be a good step toward IT conservation. I know I have too many computers running all the time (short boot up times would help!) and should think more about the energy consumption impact of my information diet.
I would have been smarter in the interview if I read How Green is my Blog on Slate by Jacob Leibenluft which does a good job of digging into the real costs of moving a gigabyte around the net. Still I wonder if all costs are being accounted for. Storage and transmission are one set of computing energy costs, how much computation do different classes of data require? The weather report takes more computation to deliver than the comics page. A web crawl and index may cost more than a hosted data set that is only minimally analyzed or recalculated. Some bytes change more often than others. Not every byte has the same amount of energy expended to get it into its current state.
