Reviews Leaderboard Database Reference Search StorageReview Discussion Reliability Survey Search About StorageReview.com Contents

Iwill SIDE-RAID100 vs. Adaptec ATA RAID 1200A
  August 14, 2001 Author: Terry Baranski  
Adaptec ATA RAID 1200A provided by Adaptec, Inc.
Iwill SIDE-RAID100 provided by Iwill Corp.


Introduction

Despite the amount of attention that high-end ATA RAID cards tend to get from many SR readers, the fact remains that it's the lower-cost adapters that make up the vast majority of the ATA RAID market. From a market share perspective, Promise's FastTrak series virtually owns the cost-sensitive ATA RAID market. This certainly doesn't mean, however, that there aren't other options for the prospective buyer.

This article takes a look at two of those options: Iwill's SIDE-RAID100, and Adaptec's ATA RAID 1200A. In our last low-cost ATA RAID comparison, it was revealed that the Promise FastTrak100 and AMI MegaRAID IDE100 perform very similarly in most tests. Was this an indication that all cost-sensitive ATA RAID cards perform about the same? Or is the story be different this time around? Only one way to find out...

The Iwill SIDE-RAID100...

Though Iwill is better known by most as a motherboard manufacturer, their SIDE-RAID series has been around since early 2000. The SIDE-RAID66 was the first member of this series, followed by the ATA-100 version known as the SIDE-RAID100. Both adapters employ ATA RAID controllers from HighPoint - the SIDE-RAID100 uses the HPT370, HighPoint's well-known ATA-100 RAID ASIC. The card has the basic features that one would expect from a cost-sensitive ATA RAID adapter, such as RAID 0, 1, and 01 support, and two ATA channels. Full specs:

  • RAID levels 0, 1, 01, and JBOD
  • Up to 4 ATA-100 drives supported (2 ATA channels)
  • 32-bit/33MHz PCI
  • Stripe size selectable from 4K to 64K
  • Hot spare capability
  • Windows 2000/98/95/NT and Linux support
  • 5 year warranty
Included with the card:
  • Two dual-connector ATA cables
  • User manual
  • Driver/utility disks

The Software...

The SIDE-100's BIOS creates and deletes arrays. Hot spares may be assigned and mirrored arrays can be rebuilt or created non-destructively. If, for whatever reason, the default disk mode (e.g., ATA-100) is undesirable, it can be changed to anything from PIO-0 to UDMA-5.

The SIDE-RAID100's software utility is very basic in nature - it simply allows the status of drives and arrays to be viewed from within Windows.

The Adaptec ATA RAID 1200A...

Recently, SR took a look at Adaptec's ATA RAID 2400A and found it to be an extremely impressive performer. Announced simultaneously with the 2400A, the ATA RAID 1200A is geared more towards the entry-level market - it's the 2400A's little brother. Like the SIDE-RAID100, the 1200A uses a single HPT370 RAID controller. Performance contrasts between the two should thus be interesting. The 1200A's full specs:

  • RAID levels 0, 1, 01, and JBOD
  • Up to 4 ATA-100 drives supported (2 ATA channels)
  • 32-bit/33MHz PCI
  • Stripe size selectable from 16K to 64K
  • Hot spare support
  • Low-profile form factor
  • Win98/Me/NT/2000 support
  • 3 year warranty
The retail kit includes the following:
  • Two dual-connector ATA cables
  • User/Installation manual
  • Driver/utility CD

The Software...

The ATA RAID 1200A's BIOS is virtually identical to that of the SIDE-RAID100, unsurprising since they both utilize the same controller.

On the software side of things, the 1200A comes with Adaptec's ATA RAID management utility. In addition to standard array monitoring functions, the utility allows arrays to be created and deleted from within Windows.

The Benchmarks...

The following benchmarks were run with a stripe size of 64K. For the SIDE-RAID100, BIOS and driver version 1.11 was used; for the ATA RAID 1200A, version 1.0 was used.

WinBench Results...

As usual, we'll start off with some single-drive scores:

Ziff Davis WinBench 99 under Windows 2000 Professional using NTFS
Benchmark AMI MegaRAID IDE100, 1 DriveAdaptec ATA RAID 1200A, 1 DriveIwill SIDE-RAID100, 1 DrivePromise FastTrak 100, 1 Drive
Business Disk WinMark 99 (KB/sec) 6304 4274 4552 6128
High-End Disk WinMark 99 (KB/sec) 15740 14760 15360 15400
Disk/Read Transfer RateStorageReview.com
Beginning (KB/sec)29800 29533 29633 29667
End (KB/sec)17500 17500 17500 17433
Disk Access Time (ms)15.18 15.74 15.32 15.14
Disk CPU Utilization (%)4.76 13.13 2.93 3.04

A couple notable results. First, the 1200A's CPU utilization comes in much higher than the rest of the pack with a score of around 13%, compared to 3% for the SIDE-RAID100 and FastTrak, and about 4.75% for the MegaRAID IDE100.

Also interesting is the 1200A's significantly higher access time - on average about 0.5ms higher.

RAID 0...

Ziff Davis WinBench 99 under Windows 2000 Professional using NTFS
Benchmark AMI MegaRAID IDE100, RAID 0, 2 DrivesAdaptec ATA RAID 1200A, RAID 0, 2 DrivesIwill SIDE-RAID100, RAID 0, 2 drivesPromise FastTrak 100, RAID 0, 2 Drives
Business Disk WinMark 99 (KB/sec) 6176 4616 4838 6124
High-End Disk WinMark 99 (KB/sec) 18100 17100 18580 9424
Disk/Read Transfer RateStorageReview.com
Beginning (KB/sec)59467 53433 58767 58700
End (KB/sec)34933 34900 34900 34667
Disk Access Time (ms)15.04 15.64 15.10 15.08
Disk CPU Utilization (%)4.70 13.30 3.00 3.03
Ziff Davis WinBench 99 under Windows 2000 Professional using NTFS
Benchmark AMI MegaRAID IDE100, RAID 0, 3 DrivesAdaptec ATA RAID 1200A, RAID 0, 3 DrivesIwill SIDE-RAID100, RAID 0, 3 DrivesPromise FastTrak 100, RAID 0, 3 Drives
Business Disk WinMark 99 (KB/sec) 6236 4652 4906 6180
High-End Disk WinMark 99 (KB/sec) 19340 17120 19140 10700
Disk/Read Transfer RateStorageReview.com
Beginning (KB/sec)75567 53967 82200 80600
End (KB/sec)52200 51800 52400 51667
Disk Access Time (ms)15.24 15.60 15.02 15.40
Disk CPU Utilization (%)4.76 13.30 2.99 3.06
Ziff Davis WinBench 99 under Windows 2000 Professional using NTFS
Benchmark AMI MegaRAID IDE100, RAID 0, 4 DrivesAdaptec ATA RAID 1200A, RAID 0, 4 DrivesIwill SIDE-RAID100, RAID 0, 4 DrivesPromise FastTrak 100, RAID 0, 4 Drives
Business Disk WinMark 99 (KB/sec) 6340 4830 5134 6356
High-End Disk WinMark 99 (KB/sec) 19200 16960 18860 8652
Disk/Read Transfer RateStorageReview.com
Beginning (KB/sec)76567 54100 82633 80667
End (KB/sec)69767 54267 69800 68467
Disk Access Time (ms)14.90 15.30 14.90 14.82
Disk CPU Utilization (%)90.43 99.00 88.67 88.70

Both adapters weigh in somewhat higher in RAID 0 Disk WinMark scores with the SIDE-RAID100 scoring higher overall. It's difficult to put much faith in these numbers, however, considering the consistent issues with RAID DiskMark scores that have cropped up over the last 6 months.

In the STR arena, it's disappointing to see the 1200A limited to around 54MB/sec. This is far lower than the competition, including the 82 MB/sec SIDE-RAID100.

RAID 1...

Ziff Davis WinBench 99 under Windows 2000 Professional using NTFS
Benchmark AMI MegaRAID IDE100, RAID 10, 4 DrivesAdaptec ATA RAID 1200A, RAID 01, 4 DrivesIwill SIDE-RAID100, RAID 01, 4 DrivesPromise FastTrak 100, RAID 01, 4 Drives
Business Disk WinMark 99 (KB/sec) 6020 4222 4630 4302
High-End Disk WinMark 99 (KB/sec) 17420 13820 17460 3080
Disk/Read Transfer RateStorageReview.com
Beginning (KB/sec)59067 53133 59233 59267
End (KB/sec)34933 34867 34900 34800
Disk Access Time (ms)15.06 15.62 15.16 13.28
Disk CPU Utilization (%)4.71 13.30 3.03 3.07

Both cards score virtually the same in RAID 1 as they do with a single drive. This isn't surprising given their low-cost nature.

RAID 01...

Ziff Davis WinBench 99 under Windows 2000 Professional using NTFS
Benchmark AMI MegaRAID IDE100, RAID 1, 2 DrivesAdaptec ATA RAID 1200A, RAID 1, 2 DrivesIwill SIDE-RAID100, RAID 1, 2 DrivesPromise FastTrak 100, RAID 1, 2 Drives
Business Disk WinMark 99 (KB/sec) 6304 4244 4478 4450
High-End Disk WinMark 99 (KB/sec) 15400 14620 15320 2920
Disk/Read Transfer RateStorageReview.com
Beginning (KB/sec)29800 29400 29800 29567
End (KB/sec)17500 17500 17500 17400
Disk Access Time (ms)15.18 15.72 15.00 13.56
Disk CPU Utilization (%)4.74 13.20 2.84 3.07

For what it's worth, the SIDE-RAID100 again wins the DiskMark battle. Unfortunately, the 1200A's higher CPU usage and access times are also present in RAID 01.

In an attempt to reduce clutter and improve readability, we've debuted graphics-based bar charts in place of the traditional HTML tables.

IOMeter Scores...

Base Scores...

Baseline with 1 Drive

In the Workstation, File Server, and Database access patterns, the SIDE-RAID100, FastTrak100, and MegaRAID IDE 100 all score virtually identical to each other. The 1200A, however, consistently lags behind by the pack by margins of 1-3%. Things become more even in the write tests, as all cards score very similarly.

Considering the 1200A's CPU utilization in WinBench, IOMeter's CPU utilization assessment should be interesting. The table below gives each card's base CPU usage in IOMeter:

IOMeter - Workstation Access Pattern - CPU Utilization (%)
Base
All LoadsLinearVery LightLightModerateHeavy
AMI MegaRAID IDE100, 1 Drive1.44 % 1.44 % 1.60 % 1.90 % 2.47 %
Adaptec ATA RAID 1200A, 1 Drive3.85 % 4.03 % 4.77 % 5.20 % 6.26 %
Iwill SIDE-RAID100, 1 Drive0.74 % 0.82 % 0.92 % 1.06 % 1.50 %
Promise FastTrak 100, 1 Drive0.87 % 0.99 % 1.09 % 1.20 % 1.68 %

The 1200A's single-drive CPU usage is significantly higher than the other cards in IOMeter as well, as is evidenced by the above numbers from the Workstation tests. CPU usage results for the other tests can be viewed in the StorageReview.com database.

RAID 0...
RAID 0 with 2 Drives
RAID 0 with 3 Drives
RAID 0 with 4 Drives

Both adapters scale similarly when moving from a single drive to a two-drive RAID 0 array. As the SIDE-RAID100 had a slight lead in the single-drive scores, it also has such a lead here. Things get a bit more interesting when moving to 3- and 4-drive arrays, however. Note that, here, the 1200A's performance is actually lower in the Workstation, File Server, and Database patterns relative to a 2-drive array. The SIDE-RAID100's performance in these tests stays about the same when scaling to 3- and 4-drive arrays, which is also what we saw with the FastTrak and MegaRAID. The 1200A's performance decrease is indeed disconcerting.

As demonstrated below, the 1200A's relatively high CPU usage continues in RAID 0:

IOMeter - Workstation Access Pattern - CPU Utilization (%)
RAID 0 with 2 Drives
All LoadsLinearVery LightLightModerateHeavy
AMI MegaRAID IDE100, RAID 0, 2 Drives1.58 % 1.56 % 1.87 % 2.31 % 3.03 %
Adaptec ATA RAID 1200A, RAID 0, 2 Drives4.67 % 4.80 % 5.21 % 6.12 % 7.54 %
Iwill SIDE-RAID100, RAID 0, 2 drives0.99 % 1.00 % 1.28 % 1.42 % 1.93 %
Promise FastTrak 100, RAID 0, 2 Drives1.05 % 1.11 % 1.22 % 1.55 % 2.10 %
RAID 0 with 3 Drives
All LoadsLinearVery LightLightModerateHeavy
AMI MegaRAID IDE100, RAID 0, 3 Drives1.54 % 1.58 % 1.87 % 2.03 % 2.97 %
Adaptec ATA RAID 1200A, RAID 0, 3 Drives4.61 % 4.65 % 5.27 % 6.19 % 7.26 %
Iwill SIDE-RAID100, RAID 0, 3 Drives0.93 % 0.99 % 1.12 % 1.23 % 1.92 %
Promise FastTrak 100, RAID 0, 3 Drives0.98 % 1.09 % 1.20 % 1.45 % 2.03 %
RAID 0 with 4 Drives
All LoadsLinearVery LightLightModerateHeavy
AMI MegaRAID IDE100, RAID 0, 4 Drives1.64 % 1.65 % 1.80 % 2.05 % 2.75 %
Adaptec ATA RAID 1200A, RAID 0, 4 Drives4.36 % 4.74 % 5.15 % 6.10 % 7.13 %
Iwill SIDE-RAID100, RAID 0, 4 Drives1.02 % 0.92 % 1.03 % 1.25 % 1.72 %
Promise FastTrak 100, RAID 0, 4 Drives1.09 % 1.09 % 1.18 % 1.47 % 2.19 %

RAID 1...
RAID 1 with 2 Drives

Unsurprisingly, both the 1200A and SIDE-RAID100 deliver RAID 1 similar to that of a single drive. It's safe to say that this is the norm for entry-level ATA RAID cards, as both the FastTrak100 and MegaRAID IDE100 also exhibit this behavior.

RAID 01...
RAID 01 with 4 Drives

Overall, the ATA RAID 1200A disappointingly performs a bit worse in RAID 01 relative to a single drive. The SIDE-RAID100, on the other hand, performs very similarly to its 2-drive RAID 0 configuration save for lower sequential write scores.

The 1200A's CPU usage is again relatively high:

IOMeter - Workstation Access Pattern - CPU Utilization (%)
RAID 01 with 4 Drives
All LoadsLinearVery LightLightModerateHeavy
AMI MegaRAID IDE100, RAID 10, 4 Drives1.73 % 1.70 % 1.97 % 2.36 % 3.07 %
Adaptec ATA RAID 1200A, RAID 01, 4 Drives4.97 % 5.23 % 5.68 % 6.81 % 8.06 %
Iwill SIDE-RAID100, RAID 01, 4 Drives1.03 % 1.01 % 1.15 % 1.42 % 2.02 %
Promise FastTrak 100, RAID 01, 4 Drives1.26 % 1.37 % 1.41 % 1.56 % 2.17 %


Conclusion...

In light of the very impressive performance of the Adaptec ATA RAID 2400A, high hopes for the 1200A may be justified. Unfortunately, the 1200A fails to impress in the performance department. IOMeter results are, overall, significantly worse than competing cards which, because of their cost-sensitive nature, don't have blazingly fast random IO performance to begin with. The 1200A's 54MB/sec STR limit is also far from impressive, especially with competing cards hovering in the 80MB/sec range.

The SIDE-RAID100, on the other hand, performs on par with the FastTrak100 and MegaRAID IDE100 overall. It's also one of the cheapest ATA RAID controllers available - attaining the retail kit for under $50 is possible. Considering the average ~$75 price tag of competing cards, this makes Iwill's adapter an attractive option for those seeking an entry-level ATA RAID card.

Contrasting the performance of the 1200A and SIDE-RAID100 demonstrates that the actual RAID controller used on a given adapter is by no means the sole determiner of performance - both of these cards use the HPT370 controller yet perform quite differently. This illustrates that there are other aspects to performance - such as drivers and firmware - that also play a role.


HOME | ARTICLES | LEADERBOARD | PERFORMANCE DATABASE | REFERENCE GUIDE
COMMUNITY | RELIABILITY SURVEY | SUPPORT SR! | ABOUT SR |

Copyright © 1998-2005 StorageReview.com, Inc. All rights reserved.
Write: Webmaster