I always thought SAN was essentially cloud storage in that you have an external pool of storage servers connected to your LAN.
By my understanding, the limitation is that traditionally, servers from different vendors in a SAN required their own specific storage because of different environments (Windows vs. Linux for example)
Does cloud storage specifically refer to being able to virtualize the servers such that it is vendor-agnostic?
In: Technology
Storage Area Networks are still devices you own at the device level, but you get to interact it with it over network protocols. So…access to the SAN can be integrated into local network topologies sometimes more cheapy, and expanding capacity can be easier as well. But, you still buy hardware, your total available disk space is the sum of your drives and so on. Cloud storage in the truest sense is either entirely or hybrid where you don’t own any hardware, it’s not located on your local network, and you probably don’t even know anything about the underlying hardware being used. In cloud storage you buy space, in SAN storage you buy devices.
traditional SAN still meant you purchased the hardware, connect it, and maintain it. it’s still
a set of physical devices sitting in one of your rooms. you need 1TB of space? you need to buy a device one time that has that much space, PO it, receive it, install it to your network, configure it….
cloud storage is provisioned without the need to figure out about hardware. you need 1TB? enter 1TB into the box and click yes. You’ll get billed for 1TB of usage monthly.
Cloud anything simply means that it’s accessible over the internet. Generally, you’re talking about a storage *service* in the case of cloud storage. Something like Google Drive or Dropbox. The service manages the storage and provides you with an interface to access your data through.
On the provider’s end, they likely are using multiple SANs to store your data. A company like Dropbox likely has a dozen racks or more of drive arrays in their datacenter, and then 20 or 30 datacenters across the world. When a new customer comes on board, they get a virtual disk that they can access and it’s copied across multiple datacenters so that their safe in the case of a natural disaster or data corruption.
But the point is that a SAN is essentially the hardware configuration, while cloud storage describes a methodology of using that hardware.
A SAN is a particular type of physical storage device (usually a big storage array connected by fiberchannel or iSCSI). SAN is synonymous with block storage, where the storage array delivers raw blocks of data to the attached computers. It can run on a dedicated storage network (fiberchannel), or over an ethernet (iSCSI)
Cloud storage is a way to buy storage, typically implying that you are renting storage from a Cloud Service Provider (like Amazon AWS, with their EBS block storage and S3 object storage). The oay-per-use model can be very good, but if you stop paying… buh-bye, data.
The cloud service provider *might* use a SAN to manage the Terabytes they rent to you, but mostly they use home-built commodity hardware (plain white-box Linux servers with the cheapest drives they can find… zillions of them, using software RAID or erasure coding to make the storage reliable)
Think of it this way: you can own a Honda Accord… that is like an on-premises SAN (though a better analogy might be a BMW because SAN tends to be high performance), or you can call an Uber which might be a Honda Accord (or a Toyota RAV4 or something else)… that is cloud storage.
For completeness, there is a NAS (network attached storage), which uses NFS or CIFS to deliver files (not blocks) over a TCP/IP network. In corporate environments the big vendor for this is Netapp, but you may have one at home if you have a router with attached storage or an Apple time jachine.
Latest Answers