James Thew - Fotolia
There's a widely held view that cloud storage will be cheaper for everything, and while it can be highly competitive, price varies depending on the use case and the particulars of what you build and whom you build it with. There are several areas that are key to look at when comparing cloud storage costs.
Performance: Storage performance varies widely based on the media type and configuration that underpins the cloud offering. If you need high performance for applications, you're likely to want storage that comes with a service-level guarantee for IOPS and possibly storage backed by solid-state drives instead of disk. If your use case is less sensitive, a "best effort" option might be good enough.
Storage type: You can get block storage for applications, object storage for files and compatible applications, file storage (with some amount of wrangling), or archival storage in the cloud. Each has variations in configurations, but the first step in understanding pricing will be to understand how you will use the storage and what's compatible with your use case and existing resources.
Capacity: There are significant discounts associated with higher levels of capacity consumption, so if you know that your usage will be high, you can expect to pay less per unit. If you're just running a proof of concept, that's fine to see how everything works, but the full-scale capacity consumption over time will make a big difference, so do your best to estimate how much space you will use in order to zero in on accurate cloud storage costs.
Durability: Many cloud providers give you the option of several levels of durability for the data you store. If the only copy of important data is stored in the cloud, you will want a high level of redundancy, but if this is a derivative copy of data stored elsewhere, reduced redundancy might be adequate, and will cost less.
Traffic: Cloud vendors charge based on the amount of each type of storage consumed, as well as for access to the data. If you are parking archival data that is unlikely to be accessed, then your traffic charges will be low, but if you're saving frequently used data, the charges can add up quickly. For many users, this is a bit of a wildcard as they aren't used to metering access to data once it's stored. Pay attention to the rules and fees associated with data traffic, as it can be complicated and vary significantly from vendor to vendor and configuration to configuration.
Time stored: With traditional on-premises storage, you pay for capacity whether it's used consistently or not. With cloud storage, you pay only when you're storing data. This can come into play for working sets of data or temporary data sets. If you need a place to put data for a short time while you work on it or move it elsewhere, the cloud can be a great fit to pay for only the duration of storage you need.
Since the price for storage varies drastically and is so different from what you have traditionally done with on-premises options, understanding your use case is critical. Start with a summary of the performance, storage type, capacity level, traffic pattern and permanence of the data that you need. You may want to experiment with high and low estimates of the variables that you are unsure of. Do a comparison of several providers for a base, as well as a high and low scenario for each variable. Using scenario comparison across several alternatives, including on-premises options, is the best way to understand the pricing impact of various storage options that achieve similar goals.
Are cloud storage price cuts adding more customers?
Price of a hybrid cloud implementation varies with definitions
Dig Deeper on Public cloud storage
Related Q&A from Andrew Reichman
With storage developments moving from hardware toward software, many vendors are reworking products to capitalize on the software-defined technology ... Continue Reading
When it comes to provisioning storage for virtual machines, techniques such as golden images and thin provisioning are key. Continue Reading
In physical environments data protection often means restoring an entire system, but using the same methodology for virtual servers is inefficient. Continue Reading