In: Computer Science
research Amazon’s in-house relational database AWS.
Consider some of the following questions.
How does Amazon approach its database solution?
- In AWS the application store the data based on its access pattern
and you predict how the data will grow further.
There are different type of data storage.
- NoSQL, Big Data, Caching, Object Storage.
How important is reliability to Amazon, and what challenges do
they face?
- reliability is an important factor which involves system to
recover any intervention or service unavailability and handling
network issues.
Any disruptions should also be handled and heavy computing should
also be managed.
What is Dynamo? What techniques does it involve?
- Dynamo is used as storage by amazon core services and they follow
key-value storage.
Technique used is vector clock to write and Merkle trees which is
used to handle temproray failure and recover permanent failure.
How does Dynamo handle failure? Does it rely on
redundancy?
- failure handling is taken care by hinted handoff, node which
detect the failure node. Another technique is comparision of Merkle
trees
of nodes and failure dectection using gossip-style protocol.
What are your personal experiences with Amazon’s
reliability?
- Its great efficient and resistent to failure.
Note : For one or two questions i didn't have much idea so i could not answer. If you still need reply back.