Amazon Web Services cloud storage options need to be strengthened

As cloud computing is widely used, the demand for cloud storage has also become widespread and diverse as one of his main trends, even in a single organization. And if you can not get the right service for your specific needs, you're likely to increase the additional costs of cloud storage. Amazon Web Services (AWS) therefore offers a wide range of options for everyone to choose from, each with a different level of performance and price. The first step in optimizing your storage service is to evaluate the strengths and weaknesses of the various AWS cloud storage options.

AWS has three storage services to choose from: Amazon Simple Storage Services (S3), Elastic Block Storage (EBS), and Amazon Glacier. Users can choose according to need a, or a variety of.

Amazon S3

Amazon Simple Storage Service is an object store designed to put large amounts of data in an orderly "bucket." This bucket is somewhat similar to a directory with a storage limit of 5TB.

Amazon S3 works better when storing content or data, frequently accessing data and being able to tolerate some performance differences, such as big data analytics and document management. S3 objects can replicate across multiple storage devices for added durability, but AWS offers de-redundant storage options with a durability guarantee of 99.99% instead of 99.999999999% of the standard S3. S3 standard storage initial price of 9.5 cents / GB / month; to redundant the initial price of 7.6 cents / GB / month.

AWS Elastic Block Store

S3 works well on programming objects, but not for some applications that require performance guarantees and access to file systems such as relational databases. EBS is better for these use cases. EBS comes with a maximum storage capacity of 1TB. Unlike S3 objects, which can be easily accessed from any device, EBS capacity is attached to a single instance of Elastic Compute Cloud (EC2). The core benefit of EBS capacity is the ability to provide I / O operations (IOPS) protection. For example, an application might need a database query that returns results in two seconds. To do this, the storage system needs to perform 1000 IOPS at the expected load. For cases that require only access to the file system, the EBS capacity is priced at 10 cents / GB / month plus 10 cents / one million IO operations. To prepare for IOPS in advance, EBS costs 12.5 cents / GB / month plus 10 cents / IOPS / month in advance. Because EBS devices are limited to 1TB, multiple EBS devices may need to be attached to an EC2 instance. It is important to note that if there are multiple IOPS pre-provisioned devices, IOPS provisioning fees will be charged for each EBS device.

Amazon Glacier

Amazon Glacier is the third option, a low-cost archive storage service. Amazon Glacier is by far the cheapest storage option, the price is 1 cent / GB / month, but with this price, there are obvious limitations.

The most important limitation is that data access operations are measured in hours rather than seconds. Because a typical retrieval operation can take place anywhere, in three to five hours, this service is only available for long-term storage of frequently accessed content. For example, Ken has used Amazon Glacier to store email and documents for regulatory compliance or e-recovery. AWS also charges additional fees for retrieving editorial data from Amazon Glacier storage within three months.

The key to optimizing AWS cloud storage is balancing costs and performance, with particular emphasis on performance, endurance, and access times. EBS storage provides guaranteed performance and file system storage, but at a higher cost than other AWS services. EBS is best for I / O-sensitive applications and can not tolerate response time differences. Providing IOPS requires an EBS-optimized EC2 instance. Instances can be thought of as specific levels of IOPS performance; for example, a large M1 instance has a cap of 500 IOPS, but a very large M1 instance has a cap of 1000 IOPS. We will find that providing IOPS requires larger machine instances, and therefore additional costs.

Evaluate the frequency of access to objects in the S3 store. If you store backups for an extended period but rarely access longer backups, these are good candidates for moving to Amazon Glacier. Storing backups in S3 makes sense at some time because it may be necessary to restore data or applications using these backups. Amazon S3 supports object lifecycle management strategies, including automating the migration of data from S3 to Glacier. Administrators can define migration policies in the Amazon Admin Console or programmatically.

【Editor's Choice】

Google Microsoft forced the palace! Amazon can keep cloud computing throne? US CIA 600 million to buy Amazon cloud services Amazon cloud storage make-up: a low profile launch of cloud synchronization features Amazon cloud computing services to lower Amazon S3 price experts explain China's cloud services market: the entire network size less than one-third of Amazon [Editor: kite Wei TEL: (010) 68476606】