I was recently reading in the New York Times “As Web Traffic Grows, Crashes Take Bigger Toll”, an article about the need for greater sustainability in Web 2.0 websites. The article provides interesting insight into the quickly growing costs of server outages for large companies like Amazon, Google, Twitter, etc.. With sites like Google Docs and Grand Central, the internet is quickly and undeniably becoming an integral part of nearly every aspect of the day. And, at the same time, there has been an increase in projects like Mozilla’s Weave and Apple’s new MobileMe, which aim to add an increased level of portability to internet use.
In coming years, the internet will no doubt play a larger role in our lives. The internet is quickly coming a greater part in previously stable services such as phone service, television service, and GPS. And, with new experiments such as smart homes coming closer to a reality and a commonality, the need for dependability is growing exponentially. Although not necessarily the main point of the article, the author raises an interesting question of what users can reasonably expect of web companies. “He [Jesse Robbins] says Web services should be held to the same standard of reliability as the older services they aim to replace.” Robbins explains this making an analogy to the lights going out at Macy’s. While this analogy makes sense for a site like Amazon, it underestimates the strength and possible pervasion of these new sites into everyday life. With sites like Grand Central and VOIP, the service being replaced currently has far better uptime than most websites; a clear challenge for these services.
To handle quick growth and increasing scaling issues, it seems that centralization and cloud computing have become staples of Web 2.0. This has become a commonality in a variety of online development forums. Hosting sites like Amazon’s EC2 and Media Temple offering server storage “in the cloud” for start ups. Web applications which offer online storage like Google Docs, Zoho, and Flickr also continue to centralize data. And lastly, services like Grand Central and Google’s App Engine have, for some, come to be necessary parts that support lifestyles and businesses. Clearly, the cloud has already begun being an integral part of our lives. However, is this for the better?
And now I come back to where I started, the article and Robbin’s explanation of how much dependability we can reasonably expect. But, is it a matter of being reasonable? As economists know, rationality is decided by observing consumers behavior, not dictated to consumers. Likewise, reasonable expectations will only be a matter of how much consumers expect, not what scientists expect of servers and dictate to be reasonable. It is for that reason that the internet originally survived. It was decentralized and as a result, an outage in one server had minimal effects on the greater internet. For this same reason, BitTorrents have flourished. And, in case Google forgot, it is what makes indexing and searching so important – it permits and even encourages decentralization content to be accessed on a far wider scale.
While movements such as these have made for some great progress online, Web 2.0 seems to have lost the essence of the internet. And, if we don’t change our ways, like the first bubble burst, the wrong falling cloud could easily seem like the sky is falling. Rather, I suggest up and coming technologies to change their ways. Technologies like Weave should not pull from the cloud all the time, but rather pull directly from other personal computers. As proven by the recent outage, Google Doc’s new use of Gears for offline use is backwards; documents should be stored on the PC and backed up online for wider use. And, hopefully, there will be an increase in home servers. Only this kind of decentralization can offer wide spread dependability, scalability, and allow the internet to continue expanding in a promsing way that can encourage it to safely play a greater role in our lives.