Welcome to MakeUseOf’s Solid State Drives 101. In this guide we’ll take a look at the the latest, greatest mass storage device and:
- Compare them to other forms of mass storage media;
- Examine how they work;
- Look at the different types available, and;
- Highlight elements of their operation and maintenance you should know.
Here is everything you’ll find in this guide.
Let’s dive in then, shall we?
1. Introduction to Solid State Drives
Computer mass storage has improved considerably since the days of punch cards (and their arch-enemies, flying insects). The computer industry survived the pains of moving back and forth through cassette tapes and swapping out floppy drives to single-write CDs. We now enjoy the convenience of large, fixed hard disk drives and rewritable optical media.
But throughout these times, one form of storage — RAM — has always been faster. Its advantage is that it accesses its data electronically, not mechanically. It suffers from one major drawback: it loses any data it holds when the machine powers down. It’s also more expensive on a per-byte basis. But a goal of the industry has long been for persistent storage to be more like RAM in terms of speed and robustness.
Developments in flash memory eventually produced Solid State Drives (SSDs), which marry the large(r) capacity and persistence of hard disc drives (HDDs) with the durability and (some of the) speed of flash memory. Let’s take a look at how SSDs came about, from both the technological and user experience viewpoints.
2. A Short History
While hard drives are obviously one parent of modern SSDs, they also owe some of their lineage to non-mechanical storage. The following all contributed, either in spirit or through technology, to today’s solid state drive.
2.1 ROM Chips
Nowadays we usually hear of Read Only Memory, or ROM(s) in the context of flashing them to a mobile device. Back before even mechanical hard drives were commonly available, computers would boot off of code stored in a ROM chip on the motherboard. As the name implies, your computer could only read from this medium. It would do so to load the initial “command prompt” from which you’d launch programs. As the capacity of these chips grew, they could hold entire operating environments like Amiga’s Workbench. They’re not quite the same as SSDs since they’re not writable, but they do represent the idea of running your OS off of non-mechanical media. Even modern EEPROMs, which you can write to, are still different since you can’t change their contents at runtime.
2.2 Hard Disk Drives/Hybrid Drives
HDDs obviously contributed to the development of SSDs, since most of the latter still use the same, or similar, hard drive standards. It made for a very easy transition from HDDs to SSDs when connection types like SATA (more on this later) make it simple to swap out an old mechanical drive for something faster. This path was made all the smoother by hybrid drives like Apple’s Fusion Drive, which added some fast solid state storage to regular hard drives.
When flash-based storage first came on the scene it was expensive. This meant it not only cost a lot, but most likely offered very little space. Manufacturers took one of two routes to marry more HDD space to the faster SSD storage:
- One approach would utilize solid state and supplement it with hard drive space. This could be a user purchasing both an SSD and hard drive, and installing programs or storing files where it made the most sense (e.g. performance-intensive apps like video editors on the better performing SSD, and less critical ones like a text editor on the HDD). There were also devices that combined both types of storage in the same physical unit, i.e. a 2.5-inch drive that housed both a mechanical HDD and solid-state storage. However these still appeared to the computer and OS as though they were two separate drives.
- Alternately, there were also more integrated examples of the above where the mechanical and solid-state storage were in the same device, but it looked like a single storage device. These devices would cache frequently-used data in the SSD portion of the drive, so it could be served to programs faster. Either the drive’s controller hardware, the OS, or both would manage what data goes where. This gave you the speed of an SSD where it could be managed, and the capacity of an HDD overall.
2.3 Flash Memory
A lot of the removable media in the 1980s through early 2000s were not so different from HDDs. Examples include writable CDs (which used optical writing instead of magnetic) to items like Iomega’s “ZIP Drive” (basically a floppy disk that approached HDD capacity). They one and all involved some sort of mechanical innards, either in the drive, the medium, or both. As portable computing really started to take hold, along with serious digital photography, more resilient forms of media appeared. Compact Flash (CF) cards were one of the first, and were joined by Sony Memory Sticks, MultiMedia Cards (MMC), Smart Media, eXtreme Digital (xD) cards, and the Secure Digital (SD) family.
These were some of the first flash-based memory, readily available for consumers, that were cost-effective and useful. They were device-dependent, however, so when the “thumb drives” that started popping up in the early 2000s, they were a significant upgrade. They were just as portable, fast enough that most users wouldn’t notice the difference, and most importantly used the ubiquitous USB port. While they couldn’t match other media in cost per MB, they represented a cheap and effective way to carry your documents around with you, for example.
2.4 Modern SSDs
A couple of factors played into the uptake of modern SSDs over HDDs or hybrid drives. Firstly, the smartphone revolution raised people’s demand for portability. After all, if you could view a full web page on your phone, why shouldn’t you be able to have a light, convenient laptop to go with it? But the memory used in those portable machines also had an effect. If your phone can hold gigabytes worth of data without heavy, loud mechanical drives or ever rising temperatures, why shouldn’t a laptop be able to do it too?
Today’s SSDs have a number of distinct advantages versus HDDs, as follows:
- Because there’s no moving, mechanical parts, when reading/writing data to a drive you’re just doing so electronically, which is very fast compared to a physical arm moving around.
- It also means that these drives are much more resistant to bumps and drops damaging them.
- The lack of moving parts eliminates the need to power those parts, saving battery life.
- The reduction in power consumption comes with a reduction in heat, which helps your machine’s performance and extends the life of your motherboard.
But how do they convey these benefits? How do SSDs actually work? In the next section we’ll take a look at the internal workings of an SSD drive, and the differences between the different models.
3. The Anatomy of the SSD
If you were to open up an SSD drive, you wouldn’t see the shiny platters and delicate actuator arm you get with a mechanical hard drive. Instead, what lies between the covers is a circuit board with a number of chips attached. There are many parts enclosed in those chips, but two that are key to understanding how SSDs work are cells and the controller. Let’s take a look at them in detail.
The basic unit of storage in an SSD is the cell. This is basically one (of many) compartments on the memory chip that contains a transistor capable of holding an electric charge. Its ability to hold this charge after power is removed sets it apart from your computer’s RAM. But otherwise, they’re similar.
Each of the cells holds a small electric charge that represents data, and when the bits are all put together they make up your files. Like hard drives though, these individual bits may live within various cells around the drive (i.e. not necessarily all contiguous cells). The same occurs in HDDs as well. But the major difference is no mechanical head needs to physically skip around the drive picking up all these bits to access a file. The SSD controller can gather quickly through electrical signals, which is why read speeds (and write speeds for that matter) are much faster compared to HDDs.
The below image illustrates how an SSD cell might store a simple text file (containing the letter “m”).
While the cells actually hold the data, the controller interprets the presence or absence of the electric charge as zero or one values. It’s also responsible for the exchange of that data with the host operating system. The controller actually houses a host of functions, as shown in the below image.
It’s sufficient for most purposes (unless your purpose is to become an electrical engineer) to know that the controller sits in between the host OS and banks of memory cells, and shuffles data between them. The above is a simplified view of how data gets from its place on the drive to your favorite application. There are lots of additional factors (like the use of pages and blocks) that contribute to the safe and efficient storage of your cat videos. A manufacturer’s controller is where they’re able to add their value, by providing unique features or the best version of those shown above.
But the memory type and controller aren’t the only factors in how well an SSD performs. We’ll dive into some of these other features and specs of solid state drives in the next section.
4. Other SSD Specs You Should Know
When considering an SSD (or a machine with one pre-installed), it’s wise to know the features and what value they provide. The following sections explain some of the main attributes of SSDs, some of the options available, and touch on some of their pros and cons.
4.1 Memory Type
In the examples above, the type of memory we’ve been describing is “flash memory,” specifically NAND Flash. NAND Flash is the same sort of memory used in removable media such as SD cards, as well as the memory in your phone or tablet. In fact, many of those mobile devices use SSDs of a particular form factor, which we’ll look at in a little bit. An alternative to this is DRAM-based SSDs, which use the same technology present in most RAM modules. While it boasts higher performance, it is very expensive, and also requires a way around DRAM’s issue of losing its contents when the power’s off. (This is accomplished by maintaining power when the machine overall is powered down, and/or with batteries.) Servers, and specifically those serving up performance-intensive applications, are the main use of DRAM. The SSDs found in the devices you’re likely to encounter are of the NAND type.
NAND Flash SSDs use one of two main types of cell formats. The first, Single-Level Cell (SLC), can only store one level of electric charge. So if the cell is charged, it equates to a value of one for a bit, and if not charged it equates to zero. Multi-Level Cell (MLC) drives can store two bits by supporting more than one charged level. A third type, Tri-Level Cell (TLC) can store three bits, although it’s not as common as SLC and MLC. Because it only has one level to support SLC is the more durable. You can effectively double your storage with the same amount of cells with MLC, so it’s cheaper. Bear in mind though, you may be writing to those cells up to twice as much and wearing the drive out faster.
4.2 Physical Interface
Another factor to consider in your SSD selection is how it’s connected to the system. In the case of mobile devices, you won’t have a choice in this matter. The manufacturer will have one of these physical interfaces as part of their design for the laptop, tablet, or phone. But you should still be aware of it, in the event that you’re weighing between devices using different interfaces. If you’re putting together your own machine, or upgrading a drive in an existing machine, you may have one or more of the following different connections available. Based on what cables or slots you have you can buy SSD that best suits your needs.
Modern hardware provides two main interfaces to storage, including SSDs:
- Serial ATA (SATA), which offers comparatively lower throughput speeds of 6 Gbps. But most motherboards will come with two or more of these, meaning you have some flexibility in designing your system. You could, for example, put your operating system and programs on an SSD while saving a larger, cheaper HDD for media storage. They’re also hot-pluggable, so if you need a blazing fast backup solution, an external SATA drive is a great choice.
- PCI Express (PCIe), which offers extremely high throughput rates of 31.5 Gbps. However, PCIe form factor is a slot on the motherboard. You may also need to forgo a dedicated graphics card, as these also use a PCIe slot. PCIe cards sub-types refer to how many “lanes” are present, where a lane is a pair of pathways, one for sending and one for receiving. The designations “x1,” “x2,” “x4,” “x8,” and “x16” indicate how many of these lanes are in the card. The image below shows a motherboard with four PCIe slots: x4, x16, x1, and x16 from top to bottom. (The very last slot is a “legacy” PCI slot.) More lanes means more simultaneous transfers, but at a cost of higher power draw when all those lanes are full. There is also a “Mini PCIe” for smaller devices.
Two other interfaces compatible with SSDs are Fibre Channel, and Serial Attached SCSI (SAS). But these are server technologies, that (for example) connect giant pools of drives together to appear as a single resource. Unless you’re a server admin you’re unlikely to encounter these.
4.3 Logical Interface
Interfaces like SATA and PCIe are physical interfaces, and define what the hardware of a cable does. Logical interfaces define a standard set of functions that storage devices offer to the operating system. They don’t specify how the manufacturer should actually make those functions happen, only that they need to be available for the device to be compliant.
Here are three common logical interfaces in use with consumer SSD systems:
- ATAPI: A much older standard developed for hard disks and other HDD-like media.
- AHCI: A newer standard for devices using the SATA bus. While it does provide some benefit to SSDs, it was designed for spinning platter-based media. It does some things less efficiently for SSDs, creating bottlenecks.
- NVMe: A standard specifically designed for SSDs. It takes advantage of properties such as multiple command queues. This means the controller can process multiple read/write commands instead of one (the case with AHCI).
This is a highly technical property of your device. What’s important for you to know is that if your device uses a PCIe-based SSD, it should also support NVMe to get the best performance.
4.4 Form Factors
A last consideration in selecting an SSD is the form factor. As with interfaces, what device you’re buying for will make a big impact on this decision. Again, if you’re looking at a tablet or handheld device, you’re stuck with what the manufacturer gives you. Laptops may or may not be the same. Some models give you easy access to the hard drive so you can replace it, in which case you’d better know what size drive you’ll need. Desktop machines may have a selection to choose from.
The PCIe models come as “cards” and not “drives,” and look not unlike other cards you may install. They’ve got a connector on the bottom, and lots of chips and diodes and stuff on the circuit-board.
The most recognizable of these will be the ones that use the same form factors as HDDs. These SATA models come in 1.8″ (second from the top in the below image) and 2.5″ (third from the top) sizes are readily available for both mobile and desktop machines. There are also 3.5″ drives (bottom) for desktops.
The top drive in the above image uses the mSATA form factor. It’s an example of devices suiting more specialized use cases, as follows:
- mSATA is a card-based format common in devices like tablets and extremely thin notebook computers. The newer M.2 form factor is an evolution of the mSATA standard, and uses more of a “blade” configuration to maximize usage of space. The above image shows an mSATA (left) and an M.2 (right) card side-by-side.
- Module-based drives package a small SSD drive into a housing designed to plug directly into the motherboard (shown in the image below). While generally available, they target specialized PCs such as those used on industrial shop floors.
Now you understand the differences between the various SSDs on the market. And maybe you’ve even purchased one, or a device that contains one. Is there anything you need to do differently? Any special processes to keep it running correctly? Let’s take a look.
5. How to Care For and Feed Your SSD
By and large, things that apply to regular hard drives also apply to SSDs in terms of making sure they run well. Try not to drop them if you can help it. Don’t get them wet. Don’t let them overheat.
But there are some special considerations to be aware of.
As we showed in the above sections, the physical construction of SSDs is quite different from older disc-based storage. As a result, some of the things required to maintain an HDD don’t apply. The most obvious of these is defragmentation. This is the process that tries to move the blocks storing the data for your files closer together. The closer together the blocks for a file are, the less the mechanical arm needs to skip around to read them and re-assemble the file in memory.
SSDs, of course, don’t have this mechanical arm. So from the performance perspective, whether or not defragging your SSD will provide any benefit is something of a question. But remember also that SSD cells can withstand only a certain number of writes before they’ll fail. That’s why several of the authors here at MUO recommend against defragmenting your SSD. Which stands to reason: if it provides doubtful gains but potentially damages the drive, why risk it?
The TRIM Command
What you should do, however, is make sure you’re using the TRIM command regularly. Like HDDs, even if you deleted a file from your OS it may still exist on the drive. It’s common that the value (charge) in that cell will remain until there’s another file that needs the space. But one unique aspect of SSDs is that they must be empty before they’re wrritten; in other words, a “rewrite” is actually a “erase-write.” If a file save operation needs to first erase all those cells before writing to them. This takes time, and results in slower perceived drive performance.
The TRIM command will identify all these “unused” cells and erase them by discharging them. In this sense it’s a kind of like the SSD’s version of “garbage collection” (getting rid of the contents of unused cells), secure delete (making those cells unrecoverable), and defrag all in one. You (or your OS) should run TRIM periodically to make drive operations faster. As a result, you should make sure you have proper support for it in your operating system, which is a convenient segue into the next section.
Operating System Support
Because SSDs connect to the machine over standard physical interfaces, modern OSes recognize them from the get-go. SATA drives should appear exactly like another hard drive. On the other hand, PCIe drives may show based on their logical interface, as is the case in Linux. As shown in the below image, my XPS 13’s disk printout clearly shows NVMe-based partitions. (These display as /dev/nvme with partition numbers afterwards).
What’s most important is whether it supports the TRIM command, and fortunately modern OSes do. TRIM support is actually a function of the filesystem, but filesystem support is dependent on the OS and its kernel. The table at the start of this section summarizes the OSes, and which filesystems compatible with those OSes, support TRIM.
SSDs Are One of the Best Upgrades for Your Laptop, Desktop, or Server
Hopefully this guide has given you a good idea of the many benefits of solid state drives over older hard disk drives. Here at MUO we’ve called an SSD “one of the best upgrades you can make” on several occasions. In particular, some reasons you might want to consider it include:
- If you have an old desktop machine that’s showing its age, consider adding an SSD. Installing the OS and programs on this drive will yield dividends in a much quicker boot time and program performance. And don’t worry, you can still use your old HDD to hold all your pictures, music, and documents.
- The above also apply to older laptops. You’ll also see an increase in battery life, and you’ll be better protected if you drop your device. Just make sure you can access your laptop’s drive bay, and that you’re buying the right form factor and interface.
- If you run a server, either at home or for a small company, an SSD will better serve your users. In particular, the multiple command queues of PCIe-based drives can attend to concurrent request to share files and/or access applications.
What do you think? Are you an “all-SSD, all the time” type of person? Or have you used an SSD to take over for an HDD that still performs “cold storage” duties? Any stories about how an SSD upgrade changed your life? We want to hear all about it in the comments below!