Cloud Layers – The network is our computer
We are comfortable and familiar with personal computing – how the device on our desk or lap offers applications (e.g. word-processing with Word); communications (e.g. email or instant messaging), storage (using the computer hard drive) and processing (using of course the computer’s CPU).
We are less comfortable and familiar with the network as our computer – how the Internet offers applications (e.g. Google Docs); communication (e.g. webmail, twitter), storage (e.g. skydrive or Amazon Simple Storage Service ) and processing (e.g. Amazon Elastic Compute Cloud ). Check the Web 2 search engine for a massive list of web applications. Check csteinberg.com for an example list of web applications that could be used in Education.
Like a personal computer the network as a computer can be regarded as set of layers – O’Reilly recently suggested three layers.
Cloud Layer 1:
As an equivalent to the physical layer he identifies “utility computing” – those offering the “lower” level foundations of the “machine” – those offering storage, computation and virtual machines – these are like Nick Carr’s “Big Switch” power stations – Microsoft’s cloud data centre even looks like a power station to me.
Cloud Layer 2:
In the second layer O’Reilly describes “Platform Services” where you can write and run applications e.g. Google App engine, and Salesforce’s force.com . Indeed, recently we had the news of the collaboration between Amazon and Force that you could run your application at Force and use Amazon for storage and database. Yahoo Pipes was an early user accessible service in this layer. Pipes allowed users to connect various data sources together and carry out limited programming functions on them. Yahoo further recognise the potential of this layer with their announcement of the Yahoo Operating System (Y!OS) concept to “rewire” yahoo, open it up for developers to access Yahoo programmatically. Since then Yahoo have been announcing an exciting set of Y!OS APIs . There is the Yahoo Application Platform (YAP), The Yahoo Query Language (YQL) and the Yahoo Social Platform (YSP). Microsoft seem to be focusing on the provision of virtual services for their server products such as Exchange, Sharepoint and SQL server in what I would regard as the layer 1 utility computing layer. However, Microsoft have a limited equivalent to pipes called Popfly that is very quick and easy to use – this flickr x live maps MASH took me less than 3 minutes to make.
Cloud Layer 3:
In the top layer O’Reilly describes “Cloud-based end-user applications” of the sort many are familiar with – Google docs, Youtube, Flickr, webmail etc.
There is another layer within the cloud which we are even less familiar with one of the fastest growing areas with the biggest potential – because it is invisible. This is the area where programs talk to each other – the Application Programming Interface (API). We are familiar with the web pages of web sites but increasingly these web sites offer a variety of ways for other sites and programs to access them.
One of the easiest ways for people to become familiar with APIs and to appreciate their potential is with the use of RSS feeds – check this video for an explanation of RSS in “plain English”. Instead of visiting a site to check for changes you can use a feed reader (e.g. netvibes, pageflakes, google reader etc) to interface with the site’s RSS feed interface. The site’s RSS feed presents site changes in chronological order and your feed reader will display these. Feed readers also allow you access and display site feeds from several sites so that you only have to access one site to view the changes from several sites (feed aggregation) – you no longer have to visit the sites themselves. You can see an example by checking my netvibes page.
In November 2008 the API tracking site the programmable web reached a much heralded milestone when it added the 1,000th API to its API directory. However, it isn’t only the quantity that is noteworthy but also the quality, variety, rate of growth, and most of all, impact that APIs are having. Most (if not all) of the major sites have APIs Google’s maps API is probably the most well known API. Flickr, Youtube and Twitter have very popular and well used APIs. Combining APIs from different sites is called MASHUP. This site presents a MASHUP of BBC news x Google maps, Gtraffic presents a MASHUP of BBC travel news and Google maps. Earth Album presents a MASHUP of Flickr and Google maps. Twittervision is a MASHUP of Twitter and Googlemaps.
The examples above are just the tip of the iceberg and like an iceberg much of the internet is invisible. Back in 2001 eBay presented one of the first APIs – this now takes over 6 billion API calls per month and accounts for 60% of eBay listings. Twitter has a rich and highly exploited API which passes 10 times the traffic of the visible Twitter site.
It is difficult to persuade development funds into invisible applications and interfaces but many are now seeing the cost advantages in developing APIs instead of the more time consuming, complex and expensive visible sites. We are seeing the development of what could be described as the “invisible web” – services on which the visible web “ecosystem” feeds. To mix metaphors, Cloud APIs could represent the larger base upon which the visible tip of the iceberg rests.
No comments yet.