Blog: Krish Krishnan

« Lifecycle of your data | Main | A closer look at UnStructured data »

A future possibility - Data Warehouse on a cloud

In the past few months there have been a number of success stories on the Amazon.com cloud computing services for web developers called AWS - Amazon Web Services which include storage database and computing platforms. At a first pass this may seem like a reinvention of the ASP hosting model in the dotcom days. The key difference here is the simple service model and the infinite scalability offered by Amazon.com.

Amazon.com’s web-based database that will work closely with its S3 online storage service and EC2 online application hosting service. S3 is shorthand for Simple Storage Service, while EC2 stands for Elastic Compute Cloud. The SimpleDB dadatabases aimed at application developers looking for a less expensive, easier-to-use alternative to a full relational database for running real-time queries against structured data.
According to Amazon.com, SimpleDB provides the core functionality of a database - real-time lookup and simple querying of structured data, without the operational complexity. The database service will automatically index data and provide performance tuning, allowing Web developers to work without the aid of a full-time database administrator.
Though its core business is an online retailer, Amazon.com's move into cloud-based computing is being welcomed by users, especially those involved in Web 2.0 computing. CoConsideringhat cloud computing is an area where Microsoft, Sun, IBM and HP also want to dominate, the are several reasons for cheer here
• The cost of the company's services is one of the benefits that have been cited. For instance, S3 lets users store 1GB of data for just 15 cents per month.
• Storing data in structured form in SimpleDB, instead of the raw storage that S3 offers, will cost $1.50 per gigabyte monthly, according to Amazon.com.
• Database requests will cost 14 cents for each machine hour of processing time used, based on the hourly capacity of a Xeon processor.
• The cost of transferring data into SimpleDB will be 10 cents per gigabyte, while outward data transfers will start at 18 cents per gigabyte for the first 10TB of information transferred in a given month.
SimpleDB is designed to store relatively small amounts of information and then to provide fast access to the data. To keep storage costs down, it recommends that users store large objects and files in S3, and put pointers and metadata associated with the files in SimpleDB.
Apart from Amazon.com, Trackvia Inc, Google and Intuit are also offering services in this area.

While this is a new technology area, it does open doors for the next generation of scalability and performance for data practitioners to think about.

Share: