Performance Comparison on Retrieving Contacts from SugarCRM with QNAP TS-112, TS-212P, and TS-119PII


I am currently porting ZurmoCRM to QTS platform which runs on QNAP device. But I have no ideas which processors is capable running database operation and there is no benchmark for reference, I invest on 3 products to test.

Test Method

I connect NAS to computer directly, install SugarCRM from App Center, and use Firefox to list all contacts. Then I use a counter to measure time to load 1, 2, and 5 contacts clicking at the same time.

I also add three different condition: enable/disable Media Library, memory_limit with different size in php.ini, and a Intel Atom D510 based Shuttle X50V2.

The key difference to QNAP: TS-112, QNAP TS-212P, and QNAP TS-119PII are processor speed (1.2GHz, 1.6GHz, and 2.0GHz), RAM size (256MB, 512MB, and 512MB), and storage (OCZ Vertex 120GB SSD, Seagate SV35 2TB as RAID 1, and OCZ Vertex 120GB SSD).

I test with 200 contact records provided from SugarCRM demo data.

Analyze

QNAP NAS Performance Comparison Loading Contacts from SugarCRM.
QNAP NAS Performance Comparison Loading Contacts from SugarCRM.

I think the database is small and will ignore the difference. Therefore, I focus on running services, software configuration, RAM, and processors.

As you may see, as a Application Server, it is a good idea to disable unnecessary service to provide more computer power. Disable Media Library bring huge improvement on performance.

The database are small, I didn’t see great different on invest more physical RAM or provide more memory for scripts. There are still lots of RAM available to use.

A better processor brings some improvement. But the improvements among Marvell 88F6281 1.2GHz (QNAP TS-112), Marvell 88F6282 1.6GHz (QNAP TS-212P), and Marvell 88F6282 2.0GHz (QNAP TS-119PII) are small, Intel Atom D510 provides much more acceptable performance for five concurrent contact loading.

Final Thoughts

For CRM, it is not a good idea to use Marvell 88F6282 SoC and Marvell 88F6281 SoC. I shouldn’t invest on these NAS and expect them to run as a micro server for CRM application. Intel Atom Processor seems to be a better choice within 5 people.

There are NAS using Intel Atom D2700 and Intel Atom CE5300 series SoC. I would like to know the performance different on CRM applications.

On the other hand, from different source and discussion, it seems that Marvell ARMADA XP provides more power than Marvell 88F6282 SoC and Marvell 88F6281 SoC.

There is any interested comparison on ARM Cortex – A8 vs. Intel Atom : Architectural and Benchmark Comparisons by Katie Roberts-Hoffman, Pawankumar Hegde. But I haven’t see any benchmark between ARMADA XP and Intel Atom Processor, I will update when I get more information.

Reference

  1. ZurmoCRM
  2. QNAP
  3. QNAP: QTS
  4. Connect NAS to your Computer Directly
  5. QNAP: App Center (QPKG)
  6. QNAP: QNAPedia Community Wiki: Category:SugarCRM
  7. Mozilla: Firefox
  8. QNAP Turbo NAS User Manual: Media Library
  9. PHP: PHP Manual: Appendices: php.ini directives: memory_limit
  10. PHP: PHP Manual: Appendices: php.ini directives
  11. Intel: Intel Atom Processor D510
  12. Shuttle X50V2 Datasheet
  13. QNAP TS-212P
  14. QNAP: TS-112
  15. QNAP TS-119PII
  16. Seagate Surveillance HDD
  17. Trusted Review: OCZ Vertex 120GB SSD review
  18. Wiki: Standard RAID levels: RAID 1
  19. Marvell: Marvell 88F6282 SoC
  20. Marvell: Marvell 88F6281 SoC with Sheeva Technology
  21. Marvell ARMADA XP
  22. Intel Atom Processor
  23. EETimes: Marvell quad-core ARM targets x86 servers
  24. SemiAccurate: Marvell ARMADA XP Processors: ARM for the datacenter?
  25. Intel: Intel Atom Processor D2700
  26. Wiki: List of Intel Atom microprocessors: CE SoCs: Dual-Core CE SoCs
  27. Intel Newsroom: Intel Launches System-on-Chip Storage Solution Designed for Simple Video Transcoding and Streaming
  28. ARM Cortex – A8 vs. Intel Atom : Architectural and Benchmark Comparisons University of Texas at Dallas EE6304 Computer Architecture Course Project – Fall 2009 by Katie Roberts-Hoffman, Pawankumar Hegde

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.