Moore’s Law states that “over the history of computing hardware, the number of transistors in a dense integrated circuit has doubled approximately every two years.” While this sentiment mainly applies to the processing power of computers, the same type of pattern can be seen in the world of data growth and storage.
The amount of data being generated today is enormous! By 2020, it’s expected to reach 40,000 exabytes (EB), which is more than 5.2 TB of data per person on the planet! That’s a LOT of data, which equates to a LOT of data storage.
Modern companies are data driven. Whether driving their own data programs or using software that is data driven (what software isn’t data-driven these days?), user data is being captured, utilized, and stored in massive scale.
Where will all this data be stored? It’s becoming inconceivable to implement and maintain on-premise solutions for data storage. Most data will live in public cloud, private cloud, pure virtualization, or a dedicated off-site solution. However you choose to offload your data, choosing the right storage solution is going to matter.
How Do You Decide Which Storage Solution(s) You Need?
As we all prepare to handle this incoming wealth of data, and as our need to store this data increases, it’s important to become literate and understand the differences in storage types, as well as the best use for each type. In this piece, we’ll focus on four “mainstream” types of data storage:
- Internal (or external) Direct Attached Storage (DAS)
- Network Attached Storage (NAS)
- Storage Area Networks (SAN)
- Object Storage
Direct Attached Storage (DAS)
- What It Is: Direct attached storage is digital storage that’s directly attached to the server that accesses the storage, as opposed to being remotely accessed through a network. Common types of DAS devices are internal/external hard drives and optical disc drives (like CD/DVD ROMs). DAS storage is a physical drive that can be directly connected to your server internally or externally, much like an external drive for your PC.
- DAS storage has a low initial cost; if you’re looking to free up space on your dedicated server and your resources are limited, a DAS device may be the perfect fit for you and your budget.
- DAS devices can offer block-level access or file-level access (typical DAS devices do one or the other), giving you the option to run applications (like email or database applications) directly from the device.
- Additionally, DAS storage is great for storing multiple types of data, especially if you don’t need super high performance.
- Some DAS systems can be expanded, giving you even more storage without having to change or swap out your existing hardware. However…
- DAS is limited to dedicated servers, meaning it’s not an option for your cloud hosting.
- While some systems can be expanded, they generally offer a smaller maximum storage capacity than other types of storage devices.
- Also, the number of servers you can connect to DAS storage is limited by how the device is connected and whether the storage is internal or external.
- In most instances (though not all), if the server the device is attached to is down for any reason, the data stored on attached DAS devices is inaccessible.
Network Attached Storage (NAS)
- What It Is: Network attached storage is an external device that’s physically attached to your server environment via LAN or WAN. Typical NAS devices are entirely self-contained, and come with their own operating system, storage, and management interface. Unlike DAS devices, which connect directly to a server, NAS devices connect through an existing network. NAS devices often come with multiple gigabit (or faster) ethernet ports to ensure high-level transfer performance.
- Because NAS devices are so self-contained, they’re extremely simple and quick to install. If you’re in a bind for space and need additional storage NOW, a NAS device can be an easy and quick fix.
- Since NAS devices are connected through a network (instead of a server), setting one up usually means no downtime. That’s always a good thing!
- NAS devices are perfect for people or businesses looking to share files over a network. As such, exports/shares can be used by multiple servers at the same time with a high number of connections.
- Because NAS devices generally run their own OS and GUI, they offer a simpler and more complete control method of controlling your data.
- A NAS is limited in what it can do. Generally speaking, they’re dedicated only to hosting files and nothing else. In that regard, it’s limited in scope.
- The performance of your NAS device is heavily dependent upon your network. You need to be very careful about deploying a NAS device in a network with limited bandwidth available.
Storage Area Network (SAN)
- What It Is: SAN is a high-speed network of storage devices that connects with servers to provide block-level storage that’s accessed by applications running on those servers. SAN lets applications (like databases) take advantage of larger, faster storage without complicated network sharing or application configuration.
- SAN setups can host storage for databases, application servers, clusters, or even file servers.
- SAN storage is perfect if you need high performance or require special feature sets (such as replication, deduplication, or extremely high availability).
- A SAN setup is inherently scalable and can grow as big as you need. There’s no limit to the number of drives you can add. (However, most systems have a finite number of drives that can be managed.)
- Because SAN systems require no reboots to add new disks, to replace disks or to configure RAID groups, you’re guaranteed 100 percent uptime.
- The high performance of SAN comes at a price: the cost per gigabyte is generally higher than it is for other types of storage.
- Because targets presented from a SAN are normally mounted on a single host, there are two major risks if you don’t have file system clustering:
- There’s a high risk of data corruption if multiple hosts connect to a SAN block device.
- Servers can’t typically read data that another server has written if they are ‘sharing’ the same block device.
- SAN generally offers no direct access to individual directories or files because storage space is presented as a block device to a server. SAN storage is specifically ordered through manufacturers/vendors, and it operates from proprietary software/firmware.
- What It Is: Object storage stores data as objects in a system, rather than as contiguous files and directories. Each “object” includes the data itself, some amount of metadata, and a globally unique identifier. Because of its unique file storage system, object storage offers an inexpensive and scalable way to store massive amounts of unstructured data.
- Object storage systems are great for bulk file storage.
- Object storage is also great for CDN, video/audio streaming, data archives, or backup targets. It’s how Spotify delivers songs and how Facebook delivers photos.
- Objects are stored redundantly, which means they’re protected. If one system goes offline, there’s no downtime — the data is still available in other places, and users aren’t affected.
- Returning to Moore’s Law and our ever-increase amounts of data, object storage is one of the best options we have to solving our increasing data growth.
- Metadata search is difficult to implement (for now, anyway).
- In most cases, you must write (or have) an application that can interact with the specific API that your object store is using, meaning you need to be really comfortable in the command line.
Storage Requirements and Considerations
Before choosing which storage solution works best for your needs, there are a handful of factors you must consider.
- Is data redundancy important/preferred/necessary? If it is, you’ll want to ensure that whatever drive you get is protected with RAID and includes multiple levels of redundancy. An object storage system would ensure that your data is kept in multiple places, ensuring maximum uptime.
- How good of a read/write performance do you need? If only the best will do, you’ll want to look at SAN storage. If you have more static storage needs, you definitely don’t need a SAN.
- How many people / how frequently will people need to access the data? This all depends upon your needs. If you’re distributing content like Spotify, object storage is a great way to ensure consistent availability. That kind of functionality is nearly impossible with a DAS.
- How much do you want to spend on your storage, in terms of cost/unit (i.e. $/GB)? If your budget is tight, you might not be able to spring for a SAN device. Instead, you could opt for a DAS or NAS option.
Each of these four storage types offers tremendous benefits in the right circumstances. There’s no hard and fast “right” or “wrong” solution — only what works best for your needs. Still not sure which storage device is right for your needs? We’re here to help. Talk to a hosting expert today!