Skip to main content Skip to Footer

Blog - Cloud ComputingCommentary from our cloud experts around the globeCommentary from our cloud experts around the globePUTINGCOMCLOUDBLOG

August 30, 2017
When using the cloud, don’t forget about storage
By: Miha Kralj

As we talk to executives about the cloud, we often encounter a common misconception: that cloud means a computing platform. That is not surprising because the computing part that carries the application is what everyone sees as the role in the cloud that is “doing” something.

However, there’s another component of the cloud that’s equally, if not more, important: storage. This is the cloud layer that holds and controls all the data that applications create and consume.

Cloud storage is critical because modern cloud architecture patterns require applications to store data in a separate persistent service rather than locally within the application. In other words, good cloud practices encourage the creation of a “stateless” architecture that prevents data from being lost or compromised in case that any part of an application or computing layer fails.

Making cloud computing decisions is easy—cloud provides relatively few possible computing options. But the cloud storage domain is much more complex; several storage and persistency choices exist. Moreover, they all have very different structure, function and usage.

Determining which is “the best” cloud storage option for your business

The fact is, selecting storage is not something you can do on a whim. It requires an independent strategic approach to cloud storage that considers trade-offs across (at least) five dimensions:

  1. Storage cost per storage unit

  2. Storage and retrieval performance

  3. Storage availability and durability

  4. Adaptability of storage structure and access

  5. Storage security

How you manage those trade-offs depends on:

  • What kind of data is in question (e.g., short-living web sessions, transactional data, or long-term archival) and how you will use it.

  • If you require super-fast performance

  • If you need a high volume of consistent streaming

  • If you’re looking to store data for a long time with little to no access to it needed

For example, let’s say you have a typical text document. If it’s something you use regularly, you’ll want to store it in a document management place like OneDrive or DropBox, which gives you immediate access to the document. On the other hand, if it’s a contract that few will ever read again but needs a place for long-term safe-keeping, a slow, limited-access archival service is best. Moreover, the difference in price points between these two options is significant.

So which type of cloud storage is best?

Let’s look at the eight main options and their pros and cons:

Storage Type Pros Cons
Disk Fast access to data and exclusive usage by each compute instance Limited growth without interruption after initial provisioning
Caching Can be shared among multiple compute instances and provides rapid access to data Requires application development knowledge to use effectively
File Provides a highly scalable, durable structure of files, can grow on demand, is easily shared Slower access time and higher latency; requires high-bandwidth network
Object Allows storage of anything of any size, access in virtually any way, and sharing Higher latency and slower access time; requires strong governance
Archiving Costs a fraction of what other storage types do Slow storage and retrieval process; requires advanced notice for retrieval
Relational Allows for easy entry, storage, query, and analysis of data—and supports complex joins Vertical scaling requires planned downtime when expanding the database
Non-relational Highly flexible, may scale horizontally on the fly, and easily supports dynamic data Requires different approach to data management architecture and lacks complex data structures
Data warehouse Can scale very quickly without the downtime issues with a relational database Isn’t as fast as relational database engines; can’t handle very complex structured data queries

The bottom line

When you create a cloud solution, you need to sort through many cloud storage and persistency services to determine which one is right for you. Selection requires more planning and consideration than most organizations initially anticipate, and certainly more than what’s necessary to make a correct cloud computing choice. When it comes to cloud storage, think carefully about your requirements. The success and longevity of your cloud solutions might depend on it.

Popular Tags

    More blogs on this topic