In fact, to build a Hadoop ecosystem is not difficult, in my understanding in difficult to make configuration adjustments based on actual demand
Traditional data warehouse, the meaning of existence is to better regulate the relationship between data, data analysis, data mining service
So do not set up an empty Hadoop ecosystem too much sense, because there is no data has no value, I do not like the reason it is because outsourcing companies outsourcing there is no way to effectively deal with some of the data independent.
ETL project meaning that you will not be much of ETL tools, but how effective engineering data extraction, data conversion, data storage. To solve the problem is to synchronize, integrity issues, whether to repeat the question, but are afraid of missing data (this would involve high concurrent processing, etc.)
Hadoop environment can provide more than the conventional environment is its efficient operation, the processing.
Project engineering, structural engineering is to take some of the thinking of
Efficient ETL process performed in parallel.