From the files of “know what someone’s title is before you post something related” comes a polite clarification from Jon Udell, who found the following quote from yesterday’s entry amusing:
Anyone who reads Jon knows that his analysis is as cogent and insightful as any analyst’s, and if I do say so myself, usually more so.
How is that funny? Check out his title.
In other news, lots of good stuff in last week’s Gillmor Gang, which featured none other than Google’s Adam Bosworth, who’s vying with Udell for the highly coveted “most mentions on tecosystems” title. One interesting tidbit from Adam was his thoughts on Audioscrobbler and the problems in information scale that the service represents. Not from a storage standpoint, but rather retrieval, manipulation, etc.
As longer term tecosystems readers may be aware, I’ve been using Audioscrobbler for a little over a month (my profile can be found here) and I’m quite fond of the service.
But what I found really interesting was what Adam didn’t really talk much about – in that example at least – which was simple application scalability. I’d guess that this is a function of the fact that Google’s got that pretty well in hand, while the data scale issues he mentions are far more complex – and lacking an adequate software solution. I do think, however, that in the shorter term simple application scalability may be more problematic, at least for players the size of Audioscrobbler. I was reminded of this as I began to examine Allconsuming.net over the past several days, as it appears to be going offline for long periods of time, indicating a difficulty in scaling.
Perhaps these scale issues are a result of precisely the data management issues Bosworth is discussing, but I think it’s equally likely that many of these applications are being subjected to loads the developers never anticipated. That problem is a lot less sexy than the one Adam’s eyeing, but I think we’ll see more of it.
Either way, one thing is clear: as more people subscribe to nework applications and services, and those services fuel more hypergrowth in content, scalability – whether it’s at the application or the data layer – will increasingly be the priority in application development.
That being the case, it would seem to give organizations that have figured out how to handle massive scale – and manage the infrastructure that requires – a decided edge. Not over the startup services like Allconsuming, Audioscrobbler, or Flickr, because I expect most of those to be acquired or cloned in the next 12-18 months, but over large competitors who might be able to meet the scale demands, but can’t do so cost effectively.
