What You Need to Know About Cloud Storage
It is that time - again. One or more of your storage units has run out of capacity, is at the end of it’s useful life cycle, has failed, or simply doesn’t satisfy your changing performance or feature requirements anymore. Wouldn’t it be nice if you could just change, add, and remove storage capacity on the fly without having to enter into significant financial and technological commitments? One way of achieving this goal is to leverage Cloud Storage. Now all you need to do is select a provider with the best price for the amount of storage you want, right? Just sign up and you’re done?
Not quite. There are a few functionally different types of cloud storage, each of them suited for different use cases. Choosing the right storage type is all about how its particular advantages serve your business goals, so in order to assess the best fit, you’ll first need to assess your needs.
Let’s look at an analogy. If you decide to buy a new vehicle, how do you choose between a sports car, SUV or pickup truck? You first need to ask yourself what you’ll be using the automobile for. Do you need to transport a lot of people? Do you want to travel very fast? Is safety a top concern? Are you looking for fuel economy? It’s the same with cloud storage: the technology is only useful in the context of what you’re trying to achieve. Let’s take a look at the different types of data storage for cloud environments, and what you need to know to make a smart decision.
What is it?
File storage means accessing single pieces of data over the network stored as files on a storage medium. To stay organized, you build a hierarchy of folders inside other folders, which contain your files. Depending on the specific variant of file storage, limitations such as access restrictions can be put on folders and files within the hierarchy.
How is it used?
File storage is best for data that is user-defined, where end users add/modify/delete the data. It’s for data that needs to be very accessible, and available quickly. Think data like Powerpoint presentations or Excel spreadsheets. File storage aims for price performance.
However, file storage does have usability limits. Because the information is so structured, scaling means not just adding space, but adding new layers to a zigzagging system. You can only have so many folders within folders before things get overly complex. Does the content of this file belong in folder x or in folder y? What if both x and y kind of fit? This system is good for maybe hundreds of thousands of files or maybe even millions of files, but not billions.
What is it?
Block storage is a very technical way of accessing file data meant to be used by applications, not people. In simplest terms, this storage type takes data and breaks it apart into chunks of information of equal size that can be accessed directly. Instead of having to follow a path of folders within folders like file storage, each block of files has its own address and contains pointers to the next block that belongs to the piece of data. You can’t find it on your own, though, as block storage is not meant for users to interact with. Since the files’ metadata - that is, built-in descriptions of what the file is, how it should be associated with other files, and who can access it - can only be read by machines, this method is only as good as the server system it employs.
How is it used?
Block storage presents the highest performance for accessing or consuming data - the type of performance required for database systems, which is the kind of environment where you’ll find it. In these kinds of systems, applications can work very quickly with block data to retrieve and modify information, faster than any human could. However, because of its performance characteristics, block storage is by far the most expensive way to approach storage.
What is it?
A third option, object storage, prioritizes efficiency and cost-effectiveness. In this methodology, all pieces of data are tagged with metadata that includes a unique ID that describes the file’s contents and location. Once all the pieces have been marked as distinct, discrete objects, they are addressed as an object – a single entity in a big, open pool of objects accessible from many different locations at once. To access the data, just use the ID or metadata associated with the object. You don’t have to worry about tiers of folders or blocks on servers because the location doesn’t impact performance. Think of it like the junk drawer in your house that has a lot of useful things inside but nothing organized in a meaningful way, yet readily available.
How is it used?
Operating like a library shelf, existing set pieces of information in object storage can be easily retrieved and used. However, the format is not meant to enable the editing and alteration of such information. This means it’s best for storing and retrieving largely static data – for example, a music mp3 file you want to listen to. The object store itself is not usually accessed directly. It is presented through an interface that accesses the object store and its distributed data behind the scenes, such as reading a book with your e-reader or watching a movie using your smart television.
Object storage is unstructured and low maintenance, so storing larger quantities of data is easy. To increase capacity, simply increase the size of the storehouse. This agility and simplicity of design provides economies of scale that impacts the associated cost.
The Right Fit
Most organizations have processes, applications, workflows, and use cases for different types of cloud storage. If you want to truly maximize productivity and minimize cost, the answer isn’t only one of the storage solutions outlined above, it’s all in the right combination. Analyzing the workload and applying the right type of storage for each one is key. Have a lot of email attachments that are accessed and shared by many people? You’re probably looking at file storage applications. Have a database or mail server? Sounds like a block storage solution. Looking to store a lot of historical financial backups data long-term? That’s a job for object storage. You need to analyze your environment to identify different types of solutions, and match the appropriate storage offering to the correct purpose.
If you need help determining which applications require which types of data storage solutions, we’re here to support you. As a trusted advisor, NFINIT is ready to help connect your IT technicalities with your end user business requirements. Contact us here.
Walter Horton, Director of Cloud and Network Infrastructure, is responsible for the infrastructure supporting NFINIT's could and network services, ensuring that they are resilient and highly available for our client's critical applications