How we test
We used IOMeter version 2006.07.27 and the N5500's performance is tested in a RAID5 'normal' state.
Setup
The host machine for IOMeter was as follows:
| Component | Details |
|---|---|
| CPU | Intel Core i7 965 Extreme (3.2GHz) |
| Motherboard | Foxconn Bloodrage X58 (P08) |
| Memory | 6GB (3x 2GB) Corsair XMS3 DDR3-1,600 |
| Disks | Seagate 500GB SATAII (ST3500320AS) |
| Graphics | Sapphire Radeon HD 4550 512MB (Catalyst 9.7) |
| Network | Realtek 8111C Gigabit |
| OS | Windows Vista Business
x64 SP1 |
Below is our IOMeter test regime:
| Option/test | Configuration |
|---|---|
| Outstanding I/Os | 10 |
| Individual test run time | 30 seconds |
| Test file | 1GB |
| Sequential read test access spec | 64KB transfers 100% sequential 100% read |
| Sequential write test access spec | 64KB transfers 100% sequential 100% write |
| Random read and write access spec | 64KB transfers 100% random 50% write, 50% read |
| 1MB Sequential read test access spec | 1MB transfers 100% sequential 100% read |
Notes
Our N5500 featured firmware version 3.00.01.
The default stripe size of 64KB was used as this corresponds with the file block size in many of the Intel IOMeter tests and should, therefore, deliver the best possible performance. We tested in two modes. The first, single Gigabit, hooks up the Thecus via, well, a single connection. The second mode, ideally for faster transfers, takes in both Gigabit connections within a load-balancing setup for the best performance.
As a comparison, we'll also be throwing in performance results for QNAP's TS-509 Pro that's run via dual Gigabit. Bear in mind that it's £100 dearer, costing £675.
