We’re starting to see a flood of 2TB drives coming on to the market from all of the major hard drive manufacturers. This review covers the Western Digital WD20EADS, a 2TB SATA-II drive with green credentials.
The WD20EADS is part of the Caviar Green range of drives. These are aimed at desktop computing and offer high capacity with power saving functionality. Physically, the drive is a standard 3.5″ form factor (1″ or 25.4mm height). The most notable physical feature of this drive is the weight. At 730g, it feels heavy and that isn’t surprising as the drive uses four platters of 500GB each to deliver the 2TB capacity. This represents an areal density of around 400Gb/in². Here’s a summary of the hardware basics:
- Model Name: WD20EADS
- Capacity: 2TB (2,000,398MB)
- Form Factor: 3.5″
- Platters: 4
- Areal Density: 400Gb/in²
- Interface: SATA 3Gb/s
- Rotational Speed: 5400-7200 rpm
- Max Sustained Throughput: 100MB/s
- Power (Typical): 7.4W
- Power (Idle): 3.9W
- Power (Standby/Sleep): 1.1W
The Caviar Green range has a number of advanced features which reduce power consumption during idle periods. These include IntelliSeek, NoTouch and IntelliPower.
- IntelliSeek – optimises seek speeds to minimise power consumption, noise and vibration
- IntelliPower – optimises spin speed (between 5400 and 7200 rpm) to reduce power consumption
- NoTouch – the reading head is doesn’t touch the recording surface.
WD claim an overall saving of 40% by implementing these saving techniques. Obviously there’s a compromise between power savings and performance, however the aim of this drive is to provide capacity rather than speed, so any penalty on performance may be acceptable. We’ll see in the performance tests.
The following graphs show performance test results of the WD20EADS using IOMETER 2006.06.27. The tests measure sequential and random I/O, covering both reads and writes. My standard evaluation PC runs Windows 7 with 4GB RAM and an AMD 64 X2 4400+ processor.
The first graph shows sequential performance with a constant queue depth as the block size is increased. With caching enabled or disabled within Windows, the throughput trends towards the expected 100MB/s as the block size increases. Compare this to the reference drive (shown in green on the graph), a WD 5000KS (500GB SATA drive), which only achieves around 70MB/s sustained throughput. This profile is pretty much what would be expected from a SATA drive; it offers good throughput for sequential workloads such as backup and archive.
The subsequent graphs show how the I/O profile changes for a fixed 16K block size and variable queue depth. For sequential I/O, the 100MB/s threshold is maintained for reads whether cache is enabled or not. Writes on the other hand, rely on the cache for improved throughput.
For random workloads, maximum throughput never exceeds 2MB/s and compared to the reference WD5000KS, the WD20EADS underperforms this model. A throughput of 1.95MB/s represents around 125 IOPS or a response time of 8ms.
The interesting comparison will be with other 2TB drives. I already have the WD2002FYPS (the WD20EADs faster brother) on test and am looking to acquire other 2TB as they come on the market.
The boundaries of HDD capacity continue to be pushed. At the current rate, we will be using 3, 4 and 5TB drives within a few years. As capacity increases continue to outpace throughput improvements, large drives (those 500GB+ today) will be used for online archiving and storage of data rather than primary input. Clearly, drives that offer power saving functionality will fare best as organisations strive to optimise large farms of multi-terabyte configurations. For the home user, storage will become an “always on” requirement, used for storing multimedia and personal information. Before we know it, the 2TB boundary will seem trivial.
This drive is now in my Drobo, giving me currently 4.5TB of raw storage capacity.
You can find the original press release for the WD20EADS here.
You can find the product details here.