How Does an SSD Work – Learning To Run With Flash

‘How Does An SSD Work’ is the fourth article in our series entitled ‘Learning To Run With Flash’.  In this series, we take the time to try to explain the ins and outs of SSDs to you the reader, as well as doing our best to provide an understanding of how this amazing technology will benefit you.  If you are new, welcome and please scroll to the bottom of this article where you will find previous articles, all compiled in a simple and progressive manner.

If you have been following along, things may seem to be a bit repetitive at times but hold on to your boots.  We are making every effort to first introduce the reader to a thought, or term, followed by later elaboration of the thought that will provide a more comprehensive look at that specific aspect of SSDs. A perfect example of this lies in our previously mentioning ‘non-volatile memory’, which you understand now is memory that doesn’t require power to retain its data.  As this has always been a key feature of hard drives, its importance can’t be understated with the SSD as well.  After all, what use would a hard drive or SSD be if you lost any information stored every time you shut the PC off, or a power loss occurred?

DATA TRAVELS LIKE A PIPELINE

Perhaps the best way of understanding data travel to or from an SSD is by drawing similarity to an oil pipeline where the flow of travel is continuous.  Regardless of how large the file is on an SSD, it’s movement to storage, or retrieval, is virtually seamless.  This characteristic is perhaps the most important quality of the SSD and differs from that of a hard drive.  Storing or retrieving HDD data is subject to maximum file size limitations and could be compared to a dump truck making several trips to a work site to bring, or remove a load of gravel.  We will provide a more detailed look at data travel of an SSD compared to that of a hard drive later, but suffice to say, this continuous flow of information affects your PC’s speed significantly.

DISK ACCESS TIME

Unlike the hard drive, SSDs have no moving parts.  Data is moved electronically in an SSD whereas, in a hard drive, spinning of the platter and placement of the HDD actuator arm are very much mechanical in nature.  Because of this, the SSD can retrieve information much faster than a hard drive, significantly reducing the disk access time. Disk access time is the length of time it takes from a request made for information, until that information is returned.

Disk access time is particularly important when it comes to executing files as it provides a visually faster experience to the PC user. Typical SSD disk access times are as fast as .01-.02ms, whereas the HDD lags along at around 9ms, some 900 times slower. Want a great example of how this translates to your experience?  If you have a computer with a hard drive, turn it on and time how long it takes before you can operate the system.  Now, compare that to an SSD which typically starts in 15-20 seconds and you just may be getting back days of your life a year from sitting and waiting for your PC to start….all because you now have an SSD!

DATA TRANSFER SPEED

Speed is the key to the SSD and that speed is measured in MB/s.  How many megabytes of data can be moved per second?  The most typical example of this might be when you move music, pictures or movies to or from your PC.  Today’s hard drives have progressed to the point that many hard drives may now be able to transfer data in the area of 100MB/s.  The typical SSD, on the other hand, can move that same data at over 5X the speed with most notebook SSDs able to reach transfer speeds of up to 550MB/s.  Several components of the SSD, such as the interface, controller and memory itself, affect the speed of data travel.

SATA 2, SATA 3 OR PCIE

One might be pretty hard pressed to find a SATA 2 SSD for sale these days and that is because the newer SATA 3 SSDs are faster and 100% backward compatible with SATA 2 PCs that so many still have.  SATA 2 relates to the interface of the SSD that is often recognized by the physical characteristics of the SATA connector, although the host interface is technically incorporated into the controller.  SATA 2 was capable of actual SSD transfer speeds up to 280MB/s and SATA 3 is capable of speeds up to 550MB/s.  Today’s newest SSDs are PCIe and not restricted to the SATA speed bottleneck, capable of well over 1GB/s transfer speeds.

SATA stands for Serial ATA and is the original bus interface that we found in storage devices.  SATA has been stepped up in revisions with SATA Rev.1 theoretically capable of 1.5Gb/s (150MB/s), Rev.2 capable of 3Gb/s (300MB/s) and the latest Rev.3 at 6Gb/s (600MB/s).  Newer SSDs also use the PCI Express bus, rather than SATA, and PCIe speeds also increase with revisions. This is beneficial as PCI Express connects directly to the CPU and offers better performance and lower latency.  PCIe Rev.1 was capable of 2.5GT/s (250MB/s), whereas PCIe 2.0 has a ‘per lane’ throughput of 5GT/s (500MB/s), and PCIe 3.0 has 8GT/s (985MB/s).  The advantage of PCIe is that we can use several lanes in a single device and an example of this is shown with the Samsung XP941 native PCIe M.2 SSD which uses 4 PCIe lanes (x4) and can reach speeds higher than 1GB/s.

UNDERSTANDING MEMORY CHANNELS

Our next step in understanding SSD speed is through explanation of SSD ‘channels’.  SSDs can be 4 or 8 channels typically and, for the most part, this is based on the controller specifications.  Whereas a hard drive has only one single path to move traffic, the SSD controller is connected to NAND flash memory (also termed flash memory or flash) by several paths (or channels), thus providing several ‘pipelines’ by which to store or retrieve data.  Internal controller instructions and firmware provide for the even distribution of data, amongst 4, 8 or 16 memory chips, all of which is indexed internally for quick storage or retrieval.

 When we speak of a SATA 3 SSD being capable of performance up to 550MB/s, this means that data can travel between the interface and controller up to that fast.  Achieving those speeds is accomplished through clever technology where more channels, accompanied by more memory chips, mean higher transfer speeds between the controller and memory.  We need to also keep in mind that this is a very general look and that the controller and memory will also have transfer speed limitations dependent on their manufacture.

HARNESSING ALL THAT SPEED

In a typical PC use scenario, the PC user will input a request onto the computer by keyboard or  mouse. The PC  will interpret those instructions and a request will be made for information from storage.  This request comes by way of the data cable from the computer to the SSD, through the SSD controller and information is then retrieved (or stored).  The speed by which that information is retrieved is determined, not only by the SSD as a whole but also, by the interface, SSD controller and memory.  Regardless of the transfer speeds that SSD is capable of, the largest visible improvement is the result of the significantly faster disk access times of the SSD executing OS files much faster than that of a hard drive.

KEEP UP WITH THE ‘LEARNING TO RUN WITH FLASH’ SERIES!

1. SLOW SSD TRANSITION AND THE CONSUMER MINDSET
2. WHAT IS AN SSD?
3. WHAT ROLE DO SSD COMPONENTS PLAY?
4. HOW DOES AN SSD WORK?
5. SSD THROUGHPUT, IOPS AND LATENCY EXPLAINED