If you have bought, or considered buying, new computer equipment in the last few years you will have come across the offering of SSD drives for storage. A high proportion of new computers and servers still use the time-tested hardware of a standard hard disk drive (HDD), but it is increasingly common to find new computer supplied with SSDs, or with a combination of SSD and HDD. So what is the difference? How do the different technologies affect cost? And when should you consider using each type of drive?
The traditional HDD utilises magnetic disks/platters and a reader on a computer-controlled arm. Like an old-fashioned record player, the disks spin and the reader detects the changes in magnetisation around each track of the platter. The disks spin at high speed. Different kinds of drives spin at different speeds, and generally, the faster the spin, the quicker data can be put on or taken off the drive. The fastest traditional HDDs run at 10,000rpm but the extra speed comes at a cost of a lower maximum capacity. Most drives currently spin at 7200rpm. HDDs are a high-capacity form of storage, with disks up to 4 terabytes in size, and except for high-end premium hardware, most drives are very cheap (larger drives are available at well under 1c/gigabyte of storage). The down-sides of HDDs are energy use, lifespan and limits on data access speed. As with any hardware with moving parts, components of a hard drive will wear out over time. HDDs use relatively high amounts of energy to spin platters at high speed, with read heads being held and moved mere nanometers above the surface. This makes them not only subject to wear and tear, but also fragile and susceptible to shocks. Sudden power failure can also damage traditional drives as the read head is left in position above the magnetic disks.
Solid-State Drives (SSDs) have no moving parts. Instead, they use flash memory – the same kind of storage found in flash cards and USB thumb-drives. Inside a standard, modern SSD you will find dozens of computer chips, comprising the digital storage of your precious documents and data. With no moving parts, SSDs are more energy-efficient, smaller, lighter and more robust than traditional HDDs. SSDs are not susceptible to sudden physical shocks as HDDs, and the loss of power will not damage them.
SSDs are also much faster than HDDs. Accessing any information from a traditional drive requires the read head to navigate to the right place on the correct platter to find the data, often multiple times within a single file if it has been split into multiple blocks of storage (fragmented). SSDs feed the data from flash storage almost instantaneously, and can do so with multiple pieces of data at a time. SSDs are often two to three times faster than even a fast HDD – to the extent that data speed is constrained by the SATA interface to the computer, not by the drive itself.
Faster, stronger, more efficient – what’s not to like? But SSDs also have disadvantages in some situations. SSDs are significantly more expensive than HDDs, currently being slightly under $1/gigabyte of storage. While they are more physically robust than HDDs, the method of operation imposes a lifespan limit on an SSD, measured in terms of the number of read/write operations. The data content on each part of the media on a flash disk can only be changed a limited number of times, usually only a few thousand. SSDs under normal operation will shift data onto new data blocks when the old ones become “tired”, but this requires extra free space to be available. The actual usable space on an SSD is thus reduced if you want to extend the life of the drive. At Adams Consulting Group, we recommend that SSDs never be used to capacity; keeping at least 20% of the drive free will mean it will last significantly longer.
Using an SSD to capacity will reduce its lifespan. Unfortunately, when an SSD develops a fault, it usually does so with little to no warning, and in many cases any content stored on the device may be unrecoverable. Vital business information should never be saved on a single SSD drive for this reason. We recommend utilising multiple drives in RAID configuration for server configuration. Employee workstations can utilise SSDs with no danger of data loss if files are stored on an appropriately configured and backed-up file server.
The cost differential combined with the limitations of read/write operations means that HDDs and SSDs each have different uses in the business environment. The ideal storage configuration combines SSDs and HDDs for different kinds of content.
- Fast-access files particularly operating system and program files – SSDs are ideal for this kind of content. You can install Windows, Office and other business productivity applications on an SSD and significantly upgrade the speed of operation of the computer, while leaving plenty of overhead on the drive to extend its lifespan. We recommend SSDs for the operating system of all workstations and servers.
- Larger data files – In most situations, data storage should be done on HDDs. Bulk storage is much cheaper in traditional HDDs, and the extra time required to access a single larger file will hardly be noticeable (as opposed to retrieving multiple smaller files, where an SSD will make a huge difference).
- Streaming files – Files designed for streaming and immediate consumption should definitely be stored on HDDs. Even a relatively slow HDD will be able to retrieve and stream the data faster than it is consumed and the large size of streaming files makes this choice an obvious one.
In some specific circumstances, for instance in a terminal services environment, it may be worth considering use of a Hybrid drive – a standard HDD with a small amount on integrated SSD storage. Hybrid drives use software to identify files that are accessed regularly, and places a copy in the SSD portion for rapid access when required. The result is a drive with the capacity and cost advantages of a traditional HDD, with the blazing speed of an SSD when it’s most useful. Hybrid drives require a period when the software “learns” the best data to cache, and are highly dependent on controller software, so they’re not as simple to manage as traditional drives, but in some situations they can be an ideal solution.
If you’re considering your storage options for your server, your network or for individual workstations, contact Adams Consulting Group for an obligation-free discussion of the most appropriate solutions.