Everything in the computing landscape is undergoing a paradigm shift, including storage. The next-generation storage system — cloud storage — is increasingly becoming a part of the enterprise IT environment in a big way. And, why not? Besides the "anywhere, anytime" data access, businesses — of all sizes — are investing in cloud storage to enjoy many computing and performance benefits that aren’t available from on premise storage; not to mention, the comparatively low total cost of ownership.
As cloud storage becomes ubiquitous in the enterprise IT space, more and more cloud storage providers are appearing to serve the market. Among them, one of the oldest, most mature, and most widely accepted is Amazon Web Services (AWS). To maximize your cloud storage investment, it is important to understand AWS storage options to make the right choice for your enterprise’s data needs.
Let us discuss the three types of storage available through AWS and best-fit use cases for any given application.
Block Storage
AWS Elastic Block Store (EBS) is AWS’s block storage option. EBS offers persistent or ephemeral disk storage that is deployed in conjunction with Elastic Compute Cloud (EC2) cloud-based virtual machines. All EBS volumes are designed for 99.999% availability, and offer durable snapshot capabilities out of the box.
AWS offers several different flavors of EBS volumes to suit the needs of your workload. At a high level, volumes fall into one of two categories: SSD-backed storage and HDD-backed storage. SSD-backed storage is best suited for databases, boot volumes, and other workloads where performance is most closely tied to IOPS. SSD-backed storage falls into two subcategories: one for general use that offers a balance between cost and performance (gp2); and the other for latency-sensitive workloads requiring specific, minimum guaranteed IOPS (io1). HDD-backed storage is best suited for throughput intensive workloads, such as MapReduce (Hadoop) and log processing, or low price-per-GB cold storage. The subcategories of HDD storage correspond to these use cases accordingly: one for throughput (st1) and one for cold storage (sc1).
Object Storage
AWS’s Simple Storage Service is the platform’s object storage solution. One of the oldest AWS services, S3 can be easily managed manually or programmatically via commonly used REST APIs and SDKs, and features excellent data durability – claimed at 99.999999999% (11 9’s) over a given year. Amazon frequently boasts this means “If you store 10,000 objects with us, on average we may lose one of them every 10 million years or so.” Even so, S3 only guarantees 99.99% availability in a given year, a key difference that was felt by many during a several hour S3 outage earlier this year on February 28, 2017.
S3 again comes in several different flavors, in this case differentiated by the frequency at which the stored data needs to be accessed. Object storage solutions can be broadly classified into “Hot”, “Cool”, and “Cold” storage. Hot storage contains data that needs to be frequently and instantly accessed, whereas Cold storage contains archival materials that are rarely accessed, but need to be retained, often for legal or regulatory compliance reasons. Cool storage falls somewhere in between, containing data with a higher likelihood of use than Cold, but less frequently used than Hot. AWS offers S3 Standard for Hot data, S3 Infrequent Access (IA) for Cool data, and S3 Glacier for Cold data. The key point for enterprises to keep in mind when storing data in S3 is the colder the storage, the less costly it will be to store, but the more costly it is to access if/when needed. Creating appropriate lifecycle policies to move data from Hot to Cool to Cold storage is a core component of maximizing the value of IT investment dollars. According to RightScale, a cloud management company, S3 Glacier beats all other providers on cold storage costs.
File Storage
Deployment of cloud-based, shared, clustered file systems is an emerging use case in the AWS ecosystem. AWS’s Elastic File System (EFS) emerged from beta in June 2016, and turns 1 year old later this month. AWS’s EFS can be mounted to multiple EC2 instances at the same time, providing a common data source for workloads and applications running across many different EC2 instances. EFS can also be mounted to on-prem instances if AWS Direct Connect is in use (high speed, private path routing into AWS), which opens up interesting use cases such as data migration to the cloud, cloud bursting, and simplified offsite backup. Pricing and sizing are completely elastic – you pay for exactly the amount of storage you consume – and the size of the file system scales automatically with no limit, while delivering 50MB per second throughput per TB.
Overall, Amazon Web Services brings a range of cloud storage solutions to the table, each optimized to meet the needs of most workloads. As the most established cloud infrastructure provider, AWS offers a suite of networking, security, and migration tools and services that allow you to completely re-platform or seamlessly extend your existing data center into the cloud where you can unlock the full potential of AWS block, object, or file storage.
Ideally, cloud storage architecture needs to be able to support both new and emerging applications such as the Internet of Things and big data analytics, as well as legacy applications and database in a reliable and scalable manner. There are many aspects of cloud storage itself that enterprises need to consider prior to choosing the appropriate service, and that is before considering the specific requirements of your organization, the size of your environment, and your budget.
F3 Technology Partners in partnership with AWS is here to help you zero in on the right cloud storage solution best suited to your organization. We will ensure you understand the options, gain a solid understanding of your expectations, needs, and limitations, and advise on the best solution for your desired results.
