Original Link: http://www.anandtech.com/show/8294/intel-ssd-pro-2500
Intel SSD Pro 2500 (240GB) Reviewby Kristian Vättö on July 30, 2014 5:00 AM EST
Last year Intel introduced the SSD Pro 1500, the company's first SSD for the business segment. The business market is different from the typical client market in the sense that businesses tend to value security and manageability over performance and price. According to a study performed by the Ponemon Institute, the average cost of a lost corporate laptop is around $50,000, which consists of lost intellectual property, data breaches and legal costs. The same study highlights that on average, 2.32% of corporate laptops are lost in one year with the total cost of lost laptops being $6.4 million per organization on average. As a result of the cost, It is obvious that companies want to minimize the damages since the value of the data is constantly increasing and so is the number of devices we carry with us.
The easiest and most efficient way to protect data is to encrypt it. While the Pro 1500 used the same SandForce SF-2281 controller as Intel's client SSDs, the Pro 1500 brought support for hardware accelerated encryption in the form of TCG Opal 1.0 compliance. The benefit of Opal is the fact that it utilizes the drive's built-in hardware encryption, which is more secure than software encryption (for instance software can always corrupt or be negated by malware) and does not degrade performance. Additionally, Opal is much easier to manage than ATA passwords because ultimately all management is done by software, meaning that all the benefits of software encryption are present (e.g. remote wiping and other IT admin features).
With the Pro 2500 Intel is updating the Opal spec to 2.0 as well as adding support for the IEEE-1667 standard, which is required for Microsoft eDrive. The Opal 2.0 spec adds some manageability features (like support for more administrators and users per device) and it also brings support for varying block sizes. Moreover, Intel is also including a new SSD Pro Administrator Tool with the Pro 2500 that brings PSID revert functionality.
Similar to the Pro 1500, Intel's vPro technology is supported in the Pro 2500. vPro and Intel's Setup and Configuration Software allow for remote management of the whole system including the encrypted drive, which makes the life of an IT administrator much easier since there is no need to physically access the device.
Another difference compared to the Pro 1500 is that the Pro 2500 will also be available at retail. The Pro 1500 was limited to OEMs and SIs, which is why we never got to review the drive. Intel said that initially it was easier to go through the OEMs to make sure the system shipped with the necessary software for Opal encryption, but nowadays the software support is much better.
|Intel SSD Pro 2500 Specifications|
|Form Factors||2.5"||2.5", M.2 2280 & M.2 2260||2.5" & M.2 2280||2.5"|
|NAND||SK Hynix 64Gbit 20nm MLC|
|4KB Random Read||24K IOPS||41K IOPS||41K IOPS||45K IOPS||48K IOPS|
|4KB Random Write||80K IOPS||80K IOPS||80K IOPS||33K IOPS||37K IOPS|
|Idle Power||5mW (DevSleep) / 55mW (Slumber)|
|Encryption||TCG Opal 2.0 + IEEE-1667 (eDrive)|
|Endurance||36.5TB (20GB/day for 5 years)|
Given that we are dealing with an SF-2281 based drive, the performance figures do not present any surprises. With compressible data even the 120GB model achieves nearly 500MB/s in sequential writes, although what is notable is that when going above 240GB the random write performance takes a rather substantial hit. This is unique to SandForce and I have seen it with other drives as well but I am not sure what exactly is causing it. I am guessing that the decrease is due to SandForce's NAND mapping table design and structure because the size of the table increases with capacities. Since SandForce does not use any external DRAM for caching of the NAND mapping table, the table size is limited and likely needs a lot more defragmentation and optimization at the higher capacities to fit in the controller's integrated SRAM cache.
Probably the most interesting tidbit of the Pro 2500 is the usage of SK Hynix NAND as this is the first time Intel is sourcing third party NAND from someone other than Micron. Bringing on SK Hynix as a new supplier will not change Intel's commitment to IMFT nor will it affect Intel's supply agreement with Micron -- it is simply an effort to guarantee sufficient NAND supply in the future. Intel already experienced some NAND shortage over the past year, so an additional NAND source is necessary as the demand for SSDs continues to grow.
Intel's contract with SK Hynix includes pre-packaged and pre-validated NAND, although Intel still does their own additional validation to ensure that the NAND meets the same quality standards as their own NAND. Intel also guaranteed that the SK Hynix NAND is SKU specific and they will not be using multiple NAND suppliers in one product. In other words, buyers do not have to beware that the NAND in the Pro 2500 or any other Intel SSD is going to change without a notice. Currently Intel is only using SK Hynix NAND in the Pro 2500 but there will be more products with SK Hynix NAND in the future.
Our 240GB sample has a total of sixteen NAND packages with each package having two 64Gbit (8GB) dies inside. Since the Pro 2500 is a SandForce drive, the capacity of one die is dedicated to RAISE parity to protect against page and block level failures.
Playing With The SSD Pro Administrator Tool
The SSD Pro Administrator Tool is a simple command line utility for IT administrators that is provided with the Pro 2500. The main function of the tool is to provide IT administrators with a way to perform a PSID revert in case the encryption key is lost. The PSID is a 32-character code that is printed on back of the SSD and basically, without the PSID and the PSID revert tool, the SSD would be a brick in case the encryption key is lost. Obiously all the data in the drive will be lost but the drive itself can be repurposed.
Unfortunately I could not get eDrive to work on my system in order to try the PSID revert because without active Opal encryption, the PSID revert command will not run at all. I really hate how difficult it is to get eDrive working because of all the requirements. With third party encryption tools, enabling Opal encryption should be as simple as flipping a switch but with eDrive that is certainly not the case (at least with a typical DIY desktop -- corporate PCs may be a different story).
Another feature in the tool is for enabling eDrive. By default, the eDrive support is not enabled (i.e. listed as 'False') but enabling it only requires a single command. The command will not actually enable the eDrive BitLocker encryption but will rather just make the drive "eDrive ready" as Intel calls it. To enable hardware accelerated BitLocker encryption, you will have to do the same steps as with any other eDrive capable drive. Intel told me that the reason behind eDrive support being disabled by default is that during the Windows install process, BitLocker encryption will automatically be enabled if the drive and the rest of the platform support eDrive (like many corporate laptops do). As many businesses use third party encryption software, having BitLocker encryption on by default would just slow down the initial install process, so Intel decided to disable eDrive and have an option to enable it if needed.
|CPU||Intel Core i5-2500K running at 3.3GHz (Turbo and EIST enabled)|
|Motherboard||AsRock Z68 Pro3|
|Chipset Drivers||Intel 18.104.22.1685 + Intel RST 10.2|
|Memory||G.Skill RipjawsX DDR3-1600 4 x 8GB (9-9-9-24)|
|Video Card||Palit GeForce GTX 770 JetStream 2GB GDDR5 (1150MHz core clock; 3505MHz GDDR5 effective)|
|Video Drivers||NVIDIA GeForce 332.21 WHQL|
|Desktop Resolution||1920 x 1080|
|OS||Windows 7 x64|
Performance consistency tells us a lot about the architecture of these SSDs and how they handle internal defragmentation. The reason we don’t have consistent IO latency with SSD is because inevitably all controllers have to do some amount of defragmentation or garbage collection in order to continue operating at high speeds. When and how an SSD decides to run its defrag or cleanup routines directly impacts the user experience as inconsistent performance results in application slowdowns.
To test IO consistency, we fill a secure erased SSD with sequential data to ensure that all user accessible LBAs have data associated with them. Next we kick off a 4KB random write workload across all LBAs at a queue depth of 32 using incompressible data. The test is run for just over half an hour and we record instantaneous IOPS every second.
We are also testing drives with added over-provisioning by limiting the LBA range. This gives us a look into the drive’s behavior with varying levels of empty space, which is frankly a more realistic approach for client workloads.
Each of the three graphs has its own purpose. The first one is of the whole duration of the test in log scale. The second and third one zoom into the beginning of steady-state operation (t=1400s) but on different scales: the second one uses log scale for easy comparison whereas the third one uses linear scale for better visualization of differences between drives. Click the buttons below each graph to switch the source data.
For more detailed description of the test and why performance consistency matters, read our original Intel SSD DC S3700 article.
|Intel SSD Pro 2500||Intel SSD 530||Intel SSD 335||Samsung SSD 840 EVO||Crucial MX100|
The SF-2281 continues to offer excellent IO consistency. It takes over 20 minutes of 4KB random writes before the Pro 2500 begins the transition to steady-state, which is slightly better compared to the SSD 530.
|Intel SSD Pro 2500||Intel SSD 530||Intel SSD 335||Samsung SSD 840 EVO||Crucial MX100|
|Intel SSD Pro 2500||Intel SSD 530||Intel SSD 335||Samsung SSD 840 EVO||Crucial MX100|
To test TRIM, I filled the drive with incompressible sequential data and proceeded with 60 minutes of incompressible 4KB random writes at queue depth of 32. I measured performance after the torture as well as after a single TRIM pass with Iometer by running a 60-second 128KB incompressible sequential write pass.
|Intel SSD Pro 2500 Resiliency - Iometer Incompressible Sequential Write|
|Clean||After Torture (60 min)||After TRIM|
|Intel SSD Pro 2500 240GB||274.9MB/s||42.2MB/s||132.0MB/s|
The TRIM issue has not changed. Again it is not a problem unless you use software encryption because otherwise there will always be compressible data, but given the Opal and eDrive support in the Pro 2500, I do not see why anyone would opt for the Pro 2500 if the plan is to utilize software encryption.
AnandTech Storage Bench 2013
Our Storage Bench 2013 focuses on worst-case multitasking and IO consistency. Similar to our earlier Storage Benches, the test is still application trace based - we record all IO requests made to a test system and play them back on the drive we are testing and run statistical analysis on the drive's responses. There are 49.8 million IO operations in total with 1583.0GB of reads and 875.6GB of writes. I'm not including the full description of the test for better readability, so make sure to read our Storage Bench 2013 introduction for the full details.
|AnandTech Storage Bench 2013 - The Destroyer|
|Photo Sync/Editing||Import images, edit, export||Adobe Photoshop CS6, Adobe Lightroom 4, Dropbox|
|Gaming||Download/install games, play games||Steam, Deus Ex, Skyrim, Starcraft 2, BioShock Infinite|
|Virtualization||Run/manage VM, use general apps inside VM||VirtualBox|
|General Productivity||Browse the web, manage local email, copy files, encrypt/decrypt files, backup system, download content, virus/malware scan||Chrome, IE10, Outlook, Windows 8, AxCrypt, uTorrent, AdAware|
|Video Playback||Copy and watch movies||Windows 8|
|Application Development||Compile projects, check out code, download code samples||Visual Studio 2012|
We are reporting two primary metrics with the Destroyer: average data rate in MB/s and average service time in microseconds. The former gives you an idea of the throughput of the drive during the time that it was running the test workload. This can be a very good indication of overall performance. What average data rate doesn't do a good job of is taking into account response time of very bursty (read: high queue depth) IO. By reporting average service time we heavily weigh latency for queued IOs. You'll note that this is a metric we have been reporting in our enterprise benchmarks for a while now. With the client tests maturing, the time was right for a little convergence.
The Pro 2500 ends up between the SSD 335 and SSD 530 in our 2013 Storage Bench. It is evident that the SF-2281 is showing its age and cannot compete with the latest high-end drives but for typical corporate workloads the performance is more than sufficient.
AnandTech Storage Bench 2011
Back in 2011 (which seems like so long ago now!), we introduced our AnandTech Storage Bench, a suite of benchmarks that took traces of real OS/application usage and played them back in a repeatable manner. The MOASB, officially called AnandTech Storage Bench 2011 - Heavy Workload, mainly focuses on peak IO performance and basic garbage collection routines. There is a lot of downloading and application installing that happens during the course of this test. Our thinking was that it's during application installs, file copies, downloading and multitasking with all of this that you can really notice performance differences between drives. The full description of the Heavy test can be found here, while the Light workload details are here.
In our 2011 Storage Benches the Pro 2500 does not do as well. The SF-2281 used to be a good drive in terms of peak IO a couple of years ago, but nowadays modern controllers and firmware dominate it.
Random Read/Write Speed
The four corners of SSD performance are as follows: random read, random write, sequential read and sequential write speed. Random accesses are generally small in size, while sequential accesses tend to be larger and thus we have the four Iometer tests we use in all of our reviews.
Our first test writes 4KB in a completely random pattern over an 8GB space of the drive to simulate the sort of random access that you'd see on an OS drive (even this is more stressful than a normal desktop user would see). We perform three concurrent IOs and run the test for 3 minutes. The results reported are in average MB/s over the entire time.
Random performance is quite typical (and unsurprising) for SandForce. Once again the performance drops when switching to incompressible data, although at low queue depths the write performance is still relatively good despite the drop.
Sequential Read/Write Speed
To measure sequential performance we run a 1 minute long 128KB sequential test over the entire span of the drive at a queue depth of 1. The results reported are in average MB/s over the entire test length.
No surprises in sequential peformance either.
AS-SSD Incompressible Sequential Read/Write Performance
The AS-SSD sequential benchmark uses incompressible data for all of its transfers. The result is a pretty big reduction in sequential write speed on SandForce based controllers.
Performance vs. Transfer Size
ATTO is a useful tool for quickly benchmarking performance across various transfer sizes. You can get the complete data set in Bench. As expected, the Pro 2500 performs similarly to the SSD 530. SandForce used to have a big advantage in write performance but those days are over.
Click for full size
Compared to the SSD 530 Intel has been able to reduce idle power consumption by nearly 20%, although the Pro 2500 is still far away from Samsung's level of efficiency. Load power consumption, on the other hand, is typical SandForce with incompressible writes resulting in high power consumption.
I have to say I have never been a fan of business SSDs. They tend to be just consumer SSDs with a couple of software features to justify the higher price tag and this is the case with the Pro 2500. Fundamentally the Pro 2500 is just an SSD 530 with TCG Opal 2.0, IEEE-1667, and vPro support. There is nothing special about the Pro 2500 and if Intel wanted, they could enable the same set of features on the SSD 530 as well. As a result, the whole SSD Pro lineup feels a bit redundant, or at least it is a very obvious effort to increase the profit margin.
I certainly understand Intel's motivation behind the separate business lineup since Intel's goal is to generate profit for shareholders, but from a consumer's perspective the model is bad. The problem is that TCG Opal 2.0 and eDrive will never become consumer friendly features if manufacturers do not include them in their client drives, which is what happens if there is a separate business lineup. Ultimately Opal and eDrive are features that benefit consumers too, even though the gains are more obvious in the business space.
|NewEgg Price Comparison (7/29/2014)|
|Intel SSD Pro 2500||$120||$160||$200|
|Intel SSD 530||$94||$140||$165|
|Samsung SSD 850 Pro||$130||-||$200|
|Samsung SSD 840 EVO||$90||-||$140|
The fact is that both Crucial and Samsung can provide the same TCG Opal 2.0 and eDrive encryption while being substantially cheaper. The 256GB MX100 comes in at almost half the price of 240GB Pro 2500, so I find it really hard to justify the extra cost. The only advantages that the Pro 2500 has are the administrator tool with PSID revert support and vPro support for remote management, although PSID revert can be done by a third party tool as well. Hence the only scenario where the Pro 2500 makes sense is a company that relies heavily on vPro for management because the Pro 2500 is the only SSD with vPro support (at least to my knowledge).
All in all, the days of Intel being an interesting player in the client SSD space seem to be over. For the past two years, Intel's focus has been in the higher profit enterprise market and that has happened at the cost of the client SSD business. Nowadays it feels like Intel is just riding on the brand they built several years ago with the X25-M series and the Pro 2500 is another example of Intel's lack of interest and innovation in the client space.