Understanding Advanced Analytics
What is advanced analytics?
Using data for future decision-making is not new but thanks to the rise of high-speed Internet connections and an improvement in technology in general companies have been able to have a large amount of data from multiple sources that they need to filter and weighing to draw conclusions and act accordingly to the information assets and patterns found.
Thanks to advanced analytics we will be able to:
- Improve our decision-making by using data more specifically.
- Automate processes saving us valuable time in recurring tasks as well as minimizing costs in this time of financial uncertainty.
- Achieve greater efficiency by focusing on processes of greater importance to our business
The role of SQreamDB
What tools do we have at hand to face the challenge?
In these times of financial uncertainty we can take advantage of the hardware we already have for bigdata with hadoop or make use of servers with the plus of nvidia video cards and with this it makes use of a tool called SqreamDB.
What is SQreamDB?
It is a tool that helps us ingesting large volumes of data from different sources in time close to real time in record processing times. Compressing the data and storing it in columnar format. We can exploit this data with optimal response times regardless of the volume of the data.
Let’s examine the benefits a bit more:
- Raw, semi-structured or structured data ingestion
- Accelerator of the processing of big data
- Processability ~ 100 times more than legacy systems
- Ability to execute queries / queries ~ 20 times faster than legacy systems
- Reduction in resource / HW costs to ~ 10% compared to legacy systems
- Ability to process large data streams in parallel using GPU video cards
- Discovery and processing of information (insights) hidden due to lack of resources
- Reduction of massive data processing times from days to hours, from hours to minutes
- Savings in storage use due to its compression system
How does SQreamDB work?
Basically it connect with any data source ingested the information fragments it and labels it from the landing zone, then compresses it and saves it. For queries to the same database, use the omission of irrelevant data or data skipping to bring the data and process it in a parallel way, obtaining better response times than hadoop or cloudera.
Can it live with Hadoop?
If you already have Hadoop you can coexist using hadoop only to extract the data from the origins and using SQream for data preparation and exploitation of it towards predictive analysis or data visualization tools.