#EC2 Analysis# This python script is for analyzing Amazon EC2 Spot Instance market historical data.
The .json files are extracted from the EC2 management website.
##Current Status## Currently, the EC2 analysis framework generates basic statistical information from Amazon's historical market pricing information.
###Stats Generated:###
- Minimum price
- Maximum price
- Mean (average) price
- Standard Deviation of price
These prices and statistics are created per-market, with a single market being defined as a specific instance type (e.g. m1.small) in a specific geographic location (VA, CA, or Ireland).
##Next Steps## After loading all of the JSON historical data into a SQLite database, I intend to hook a python daemon up to Amazon's EC2 api, and begin continuously populating the pricing information.
I also want to be able to do time-of-day queries and other specific statistical regressions on the data set.
###Amazon JSON structure###
{"requests":
{"describeSpotHistory":"36df7f6c-58f6-4079-816a-475b5a08080b"},
"priceChange":[
{"instanceType":"m1.small",
"productDescription":"Linux/UNIX",
"spotPrice":0.029,
"timestamp":"2010-04-17T06:11:24.000Z",
"epochtime":1271484684000},
{"instanceType":"m1.small"
,"productDescription":"Linux/UNIX",
"spotPrice":0.031,
"timestamp":"2010-04-17T04:52:45.000Z",
"epochtime":1271479965000}
]
}
###SQLite Table schema:### CREATE TABLE pricehistory ( location TEXT, instanceType TEXT, spotPrice DECIMAL, timestamp TEXT )