Everyday Reflections

Good

• The ability to provision huge databases as needed, without going through a costly and slow procurement process to obtain the hardware and software
• The ability to scale to handle huge databases, perhaps well beyond the petabyte range
• The potential to use an elastic set of resources to return result sets with enough speed to be actually relevant when operating a business
• The potential to save huge amounts of money over the years versus the cost of using your own hardware and software
• Built and optimized for doing aggregation queries over large sets of data. When we want to answer a question with Redshift, we just write a SQL query and get an answer within a few minutes—if not seconds.
• From a user perspective , any one of our developers can write a SQL query and they have an answer to their question in less than 5 minutes. Moving from even a Hadoop based workflow to an interactive console session with Redshift is a major improvement. I assume moving from RDBMS to this would definitely be a big one !!
• Additionally, since much of the user facing bits of Redshift are based on PostgreSQL there is a large ecosystem of mature, well-documented tools and libraries for us to take advantage of.
• Impressive web management console Amazon provides with Redshift. For a 1.0 product, the console is comprehensive and offers much more information than we expected it to.

Bad

• The possibility of outages; it’s not that your internal data warehouse does not go down at times, but any failures will be public and give cloud computing a black eye internally
• The costs of data migration and integration; in many instances, you’ll need huge amounts of bandwidth to transmit the data from internal systems to the cloud-hosted Redshift, or you’ll be shipping USB drives via FedEx to Amazon Web Services
• A lack of best practices; we just started with public cloud-hosted data warehouses and clearly have some things to learn
• The possibility of higher costs; although many organizations will find cost savings with cloud-hosted databases such as Redshift, many will discover that their cloud computing bill is much higher than anticipated — perhaps exceeding the cost of an on-premise database
• Security issues with public cloud and data leakages to that effect
• Depending on from which instance this is accessed it will be through internet traffic as opposed to MPLS hence it can be slow
• Write performance of Amazon Redshift is relatively low compared to „classical“ relational databases (in your data center) as you have to upload all data into the cloud
• High Availability
• Full table scan
• Data Loading

Very Good Reference: 

http://word.bitly.com/post/48854093418/speeding-things-up-with-redshift

Summary from above:

Can speed up and expand our ad hoc data analysis.