Kinaxis - RapidResponse: How is it so fast?


Published on

Here's a presentation by Ravi Puvan, senior product manager, Kinaxis, on how RapidResponse is so fast compared to other supply chain software on the market.

To view the full technical whitepaper, visit:

Published in: Business, Technology
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Kinaxis - RapidResponse: How is it so fast?

  1. 1.
  2. 2. By Ravi Puvan, senior product manager, Kinaxis<br />RapidResponse: How is it so fast? <br />
  3. 3. RapidResponse: How is it so fast?<br />Kinaxis has built specialized supply chain software expertise and experience over many years.<br />RapidResponse is unrivaled in speed and performance breakthroughs. <br />This PPT will show the technology architecture and analytic capabilities of RapidResponse. <br />
  4. 4. Technology Architecture<br />RapidResponse uses an in-memory database.<br />Relies on main memory for computer data storage. <br />Employs patented technology to store multiple versions of data using only incremental changes (deltas) in input date.<br />Provides access to numerous scenarios without incurring large storage costs.<br />RapidResponse<br />Scenario B<br />Base Scenario <br />Scenario A<br />
  5. 5. Technology Architecture<br />Competitive systems store complete copies of each set of input data. <br />However, the storage requirements to save complete copies of each dataset (or version) can be very large, as shown below. <br />Other Database<br />Scenario A<br />Base Scenario<br />Scenario B<br />
  6. 6. Database Design<br />Dealing with complex data relationships like the Bill of Materials (BOM), the database design employed by RapidResponse provides a significant performance advantage over other competing designs like relational and columnar databases.<br />Comparing Database Design<br />Poor<br />Excellent<br />
  7. 7. Relational Database<br />Pros:<br />Most common choice for data warehouse applications.<br />Data is organized in tables and allows for query flexibility. <br />There is a vast array of analytic and integration tools built for relational databases. <br />Cons: <br />Analytic-intensive applications are difficult to design, challenging to maintain, and require designers to compromise between query performance and maximizing query flexibility. <br />Performance breaks down when handling complex relationships like BOMs.<br />
  8. 8. Columnar Database<br />Pros: <br />Stores content by column instead of row<br />Can reduce the storage requirements where there is data duplication. <br />Advanced search functions<br />Spans fewer columns of aggregated data<br />Cons:<br />Can’t handle complex relationships like BOMs.<br />Less efficient when many columns of a single row are required at the same time. <br />
  9. 9. RapidResponse Database<br />RapidResponse uses direct memory references to quickly access the necessary data.<br />Speed is significantly improved when accessing data. <br />Independent Demand<br />Ref<br />Demand 1 for Part ABC<br />Part<br />Ref<br />Demand 2 for Part ABC<br />Set<br />Part ABC<br />Ref<br />Demand 3 for Part ABC<br />Demand 4 for Part ABC<br />Ref<br />
  10. 10. RapidResponse Optimized Analytics <br />Analytics code is directly compiled into the database engine. Has direct access to the memory data.<br />Less moving of data between database and analytics improves performance. <br />Analytic Modifiers <br />Introduces additional logic to be considered by core analytics. <br />Can then modify calculations with additional logic. <br />
  11. 11. RapidResponse Analytics Examples<br />Capable-to-Promise<br />Uses patented method to calculate realistic order completion dates. <br />Following calculations are performed by the CTP analytics:<br />Supply demand allocation.<br />Available date.<br />Gating part.<br />Incremental availability.<br />Multi-Sourcing<br />Allows a part to have more than one source of supply to satisfy demand requirement. <br />Allows for allocation of different sources of supply based on factors such as; target, priority, and contractual agreements. <br />
  12. 12. RapidResponse Performance Tests<br />
  13. 13. RapidResponse Performance Tests<br />Simple Query Speeds<br />Add up all the quantity fields on the scheduled receipts: 0.050 sec <br />Count number of scheduled receipts whose start date is more than 90 “working days” before the due date: 0.065 sec<br />Produced report of parts sorted by site and name: 7.0 sec<br />Complex Query Speed <br />Run complete netting (MRP calculation) and count the number of planned order recommendations: 45 seconds!<br />Running the same calculation again: 5 seconds!<br />
  14. 14. You can download the full technical whitepaper “RapidResponse: How is it so fast?” at:<br />
  15. 15. Thank you! <br />Copyright © 2011 Kinaxis Inc. All Rights Reserved.<br />