Questo PDF è disponibile solo per il download

Configuring an In-Memory BI Platform for Extreme Performance

To deliver the extreme query responsiveness required for real-time analysis of high-volume data sets, Intel IT conducted tests to find the optimal platform for a cost-effective, high-performance in-memory business intelligence (BI) solution. Performing server-sizing and stress tests helped us find the best price/performance combination of server speed, number of processor cores, cache size, and memory for industry-standard servers based on the Intel® Xeon® processor family.

In-memory BI solutions provide the enhanced access and response capabilities organizations need to deliver the right information to the right decision makers at the right time. Since in-memory BI solutions differ in many ways, we performed tests to determine the best platform for our selected third-party in-memory analytics application.

By determining the optimal configuration for our in-memory platform, we expect to achieve the following advantages:
• Cost-effective performance in an enterprise-class data warehouse built for our in-memory BI solution
• High business value through a solution that enables Intel business groups to achieve real-time visibility into high-volume data sets, faster time to insight, shorter development times, and new self-service BI opportunities
• The ability to easily scale and replicate our solution for future applications of in-memory BI solutions

The exceptional velocity and sub-second latency of an in-memory database is becoming important to Intel IT’s overall multiple data warehouse strategy as a way to deliver more powerful analytics capabilities to business groups across Intel. In the near future, we may deploy in-memory BI solutions to big data use cases such as supply-and-demand planning, near real-time identification of new business opportunities, balance-sheet hedging that potentially saves USD millions in foreign currency translation, and real-time supply-chain risk assessment on more than 400,000 parts and USD 1.6 billion in expenditures.

To deliver the extreme query responsiveness required for real-time analysis of high-volume data sets, Intel IT conducted tests to find the optimal platform for a cost-effective, high-performance in-memory business intelligence (BI) solution. Performing server-sizing and stress tests helped us find the best price/performance combination of server speed, number of processor cores, cache size, and memory for industry-standard servers based on the Intel® Xeon® processor family.

In-memory BI solutions provide the enhanced access and response capabilities organizations need to deliver the right information to the right decision makers at the right time. Since in-memory BI solutions differ in many ways, we performed tests to determine the best platform for our selected third-party in-memory analytics application.

By determining the optimal configuration for our in-memory platform, we expect to achieve the following advantages:
• Cost-effective performance in an enterprise-class data warehouse built for our in-memory BI solution
• High business value through a solution that enables Intel business groups to achieve real-time visibility into high-volume data sets, faster time to insight, shorter development times, and new self-service BI opportunities
• The ability to easily scale and replicate our solution for future applications of in-memory BI solutions

The exceptional velocity and sub-second latency of an in-memory database is becoming important to Intel IT’s overall multiple data warehouse strategy as a way to deliver more powerful analytics capabilities to business groups across Intel. In the near future, we may deploy in-memory BI solutions to big data use cases such as supply-and-demand planning, near real-time identification of new business opportunities, balance-sheet hedging that potentially saves USD millions in foreign currency translation, and real-time supply-chain risk assessment on more than 400,000 parts and USD 1.6 billion in expenditures.

Video correlati