Apptio's release of Amazon S3 Rightsizing Recommendations provides visibility into bucket-level metrics and surfaces rightsizing recommendations to uncover cost savings opportunities within their S3 spend.
About Amazon S3
Amazon Simple Storage Service (S3) is a popular cloud object storage service used for common enterprise storage needs – images and videos, static website content, big data analytics, backups, and file sharing, among many others. This AWS service typically represents the third- or fourth-largest cloud spend category across Apptio Cloudability customers, including virtual machines (Amazon EC2), persistent disks (Amazon EBS), and relational databases (RDS), and accounts for around 10-15% of total AWS spend.
Buckets, objects, and storage classes are key concepts to know in Amazon S3. Amazon S3 is organized around buckets, which are the top-level containers. Inside these buckets reside the individual objects (consisting of object data and metadata), which can additionally be organized hierarchically within folders. Storage classes are tailored based on access patterns, ranging from frequent to infrequent (archival) storage. It is important to note that an object, not a bucket, is associated with a storage class.
Here are the currently available S3 storage classes, each differentiated by price, performance, availability, minimum size, and duration.
Organizations can have up to one thousand buckets per account . Each of these containers might contain a few to several hundred thousand (even millions) individual objects, and up to five terabytes of storage per bucket. Costs are incurred at the rate for the object’s storage class and factored by the allocated storage space (GB/month), duration, number of operations (read, write, lists, delete), and volume of data transferred (egress). Depending on the storage class, additional fees might be incurred, such as monitoring, storage class transitioning, and restoration fees.
By default, Amazon S3 objects are created in the highest-priced storage class — Amazon S3 Standard. In fact, our analysis of Cloudability customers shows that over 93% of all objects belong to this storage class. To manually optimize objects, you need to identify the rates for each storage class by region, collect historical access patterns and operations, then evaluate each of the alternatives, considering all the trade-offs. Now scale this by hundreds of buckets and millions of objects and it’s easy to understand how Amazon S3 optimization presents a complex challenge for many organizations.
Insights and optimization
Amazon S3 Rightsizing Recommendations offers insights and visibility into critical bucket-level attributes and metrics — storage class, storage size, count of objects, and operations. Furthermore, in the case where a bucket consists of objects of heterogeneous storage classes, this feature provides the breakout of these classes. Based on our rightsizing model, recommendations detail the optimal storage class, risk, and estimated cost savings of acting against that opportunity. Our analysis has identified up to 25-35% saving opportunities are available by following these recommendations.
Amazon S3 Rightsizing recommendations list
Below is the Amazon S3 Rightsizing Recommendations list. Refreshed daily, the Cloudability platform analyzes your Amazon S3 cost, usage, and utilization metrics looking back the past 30 days to produce bucket-level insights and optimization recommendations.
Here is a snapshot of the Details Pane. This example highlights a mixed storage class bucket (comprised of three different storage classes) and details bucket-level metrics including storage size, read/write request count, count of objects, data transferred, data retrieved, and early deletion.
To learn more, sign up for a Cloudability free trial today!