Top 4 Considerations When Evaluating a Data Analytics Platform

From fraud detection to clickstream analytics to simply building better products or delivering a more optimal customer experience, Big Data use cases are abounding with analytics at the core.

With a solid business or use case in place, the next step that organizations typically take is to investigate and evaluate the appropriate set of analytics technology from which to accomplish their analysis, often starting with a data analytics platform. But what are the requirements from which to base your evaluation?

The Winter Corporation, the large-scale data experts, just finalized an in-depth white paper (The HP Vertica Analytics Platform: Large Scale Use and Advanced Analytics) that reflects the results and findings through evaluation, independent research, customer and employee interviews, and documentation review.

Intended for a more technical audience, this white paper focuses on key evaluation criteria that your organization can use as a guide as you conduct your own evaluation.



Winter Corporation identified these key feature areas as critical for any data analytics platform:

1. Architecture
• Column store architecture
• Shared nothing parallelism
• Cluster size and elasticity
• Smart K-Safety based availability
• Hybrid storage model
• Multiple database isolation modes
• Both bulk load and trickle feed

2. Performance
• Extensive data compression and data encoding
• Read-optimized storage
• Highly parallel operation
• Storage of multiple projections
• Automatic physical database design

3. General Useful and Noteworthy Features for Large-Scale Use
• Export-import
• Backup/restore
• Workload analyzer
• Workload management
• Role-based security

4. Extensions for Advanced Analytics
• SQL extensions
• Built-in functions
• User-defined extensions
• Flexibility in accessing and analyzing all data (structured, semistructured, or unstructured)

Finally, once you have evaluated and confirmed that the data analytics platform meets your feature and technology requirements, you want to hear from other organizations that have deployed large-scale analytics’ initiatives in real-world environments.

The white paper concludes with a write-up on how Zynga, a social game services company with more than 240 million users of its online games, stores the actions of every player in every game — about 6 TB per day of data — in near-real time in the HP Vertica Analytics Platform. No matter where in the world a game event occurs, the data can be retrieved via a report or query from the central HP Vertica database no more than five minutes later.

Get Started With Vertica Today

Subscribe to Vertica