Shulou(Shulou.com)06/02 Report--

What is serverless storage / serverless database? How is it different from other storage / database services? This article will introduce the basic properties of serverless storage services and illustrate the differences between different serverless storage and serverless databases through some use cases in serverless computing. Finally, we will continue to discuss serverless storage services that have not yet emerged, and look forward to the future of this field.

Serverless means no server?

Source-http://www.commitstrip.com/en/2017/04/26/servers-there-are-no-servers-here/

The term "serverless" is not really appropriate. No server does not mean that the server is no longer needed, but adopts the concept of abstracting the technology away from the user, so that the user no longer has to worry about the server, and others (cloud providers) will help the user with the relevant work.

Serverless computing

Nowadays, serverless is a very popular word. Although the term "serverless" has been around for some time, the main reason it has become so popular is the introduction of AWS Lambda in 2014. With the launch of Amazon API Gateway in July 2015, serverless has become more popular. In general, serverless (or serverless computing) refers to applications whose server-side logic is written by developers and runs in stateless computing containers-- containers that are event-triggered, short-lived (perhaps only one call), and fully managed by a third party. This is also known as function as a Service (FaaS).

Stateless nature of serverless computing and storage requirements

One feature of serverless computing or FaaS is that it is stateless. Therefore, it is not possible to maintain a state between two executions of the same function in a general design. This is why serverless storage is introduced, and if you need to maintain state between two FaaS functions, you should store it in an external storage service.

Problems with traditional Stora

Traditional storage solutions are designed to run continuously on a set of fixed servers in a single geographic location, and developers need to know a lot about configuration details to use these storage solutions correctly. In the worst case, the pros and cons of high availability and performance depend on what the developer knows inside the database. Developers need to figure out and configure areas, space, volumes, memory, computing capacity, and software versions. The development team must spend hours thinking about capacity planning, configuration, sharding, backup, performance tuning, and monitoring. This is a bit inappropriate when you need to spend 80% of your time setting up and operating the database, and only to support serverless functionality. Although serverless computing adds flexibility to the computing layer, you may not be able to give full play to its advantages when the storage layer does not provide the same resilience.

True serverless storage

Ideally, truly serverless storage should have the following properties:

No configuration required

Truly resilient (vertical scale-up without operator intervention)

Pay on demand

When using serverless storage, developers do not have to worry about any infrastructure details, such as node size, memory, or storage size. In the same way that AWS Lambda charges for a single function call, resource consumption in serverless storage should be measured according to the amount of calculation and storage used by different workloads. This is the essence of serverless databases, where users never pay for idle capacity, and storage usage reflects function calls and workloads.

With traditional storage products, overprovisioning is the only strategy to prepare for peak traffic, and the resilience of serverless storage prevents overconfiguration that can lead to waste. Because serverless storage can scale flexibly without user intervention, developers can start without capacity planning, and applications always have sufficient capacity. In addition, the pay-on-demand pricing model will ensure that there are no idle resources to increase the cost of users, the cost of serverless storage only expands with use, and you will never pay for unused capacity.

No server storage optional

Today, there are multiple serverless storage categories, and there are usually multiple cloud providers offering more or less the same services. In this section, we will introduce these different serverless storage categories, as well as examples of each category and some of their use cases.

Serverless object storage

Object storage is a non-hierarchical method of data storage, which is usually used in the cloud. Unlike other data storage methods, object-based storage does not use directory trees. Discrete data units (objects) exist at the same level in the storage pool. Each object has a unique identity name that the application uses to retrieve it. In addition, each object may have metadata for retrieval. Some common examples of serverless object storage are Amazon S3 Google Cloud Storage, Azure Storage and IBM Cloud Storage. A common example of using serverless object storage in a serverless computing environment is thumbnail generation.

1. The mobile application uploads images to the object library.

2. Serverless object storage triggers a change event after uploading the image, thus performing serverless functions.

3. The serverless function creates thumbnails based on uploaded images.

4. After uploading the thumbnail to the serverless storage, the serverless function sends a push notification to the mobile application.

5. The mobile application downloads the thumbnail and updates the user interface accordingly.

Serverless relational database

An example of a serverless relational database is Amazon Aurora Serverless, which was launched in the last quarter of 2017. It has two different versions compatible with MySQL and PostgreSQL, but it is also compatible with other known systems such as MariaDB,Oracle. Amazon Aurora Serverless is a database service that automatically expands on demand, and the database automatically starts, shuts down, and expands capacity according to the needs of the application. It enables users to run their databases in the cloud without actually managing any database instances. It is a simple, cost-effective choice for uncommon, intermittent, or unpredictable workloads, so it is ideal for serverless computing.

Serverless NoSQL database

Serverless Key-value and document storage

There are several different types of serverless NoSQL databases, the most common of which are Key-value storage and document storage. The Key-value database is an example of data storage for storing, retrieving, and managing associative arrays, a data structure now commonly referred to as a dictionary or hash table. A dictionary contains a set of objects or records, which in turn contain many different fields, each containing data. These objects are stored and retrieved using a Key that uniquely identifies the record, and are used to quickly find data in the database. Document-oriented database or document storage is used to store, retrieve and manage document-oriented information, also known as semi-structured data. To give a few examples: Amazon DynamoDB,Google Cloud Datastore and IBM Cloudant NoSQL DB. Earlier, Amazon announced the launch of DynamoDB on-demand, which provides on-demand read / write capacity options for pay-on-demand DynamoDB. This serverless Key-value and document storage can be used for a variety of purposes, and one simple use case is to use serverless functionality to process contact us (Contact Us) forms.

Serverless time series database

Time series database (TSDB) is a software system optimized for processing time series data, which is an array of numbers indexed by time (date-time or date-time range). Time series data has specific characteristics, for example, it usually arrives in the form of time sequence, the data is Append-only only, and the query is always in a time interval. Although relational databases can store this data, they are inefficient in processing it because they lack optimization, such as storing and retrieving data at intervals. An example of a serverless time series database is the recently published Amazon Timestream, which is a dedicated time series database that can effectively store and process records at time intervals.

Source-https://aws.amazon.com/timestream/

Serverless graph database

Quite simply, the design of the graph database (Graph Database) treats the relationship between the data as important as the data itself, and does not limit it to a predefined model when saving the data. Instead, the data is stored as if it were drawn-- showing how each entity is connected or related to other entities. FaunaDB is a serverless graph database. It is a general, transactional, temporal, geographically distributed, strongly consistent and relational database. FaunaDB also supports document storage and time series database functions.

Looking to the future, what is lacking?

Serverless memory data storage / caching

One of the storage types that are not yet available in serverless storage is in-memory data storage. One of the options we currently use for caching is caching in API Gateway. Amazon DynamoDB has a unit millisecond delay, and if the user needs to reduce the latency slightly, you can use the DynamoDB Accelerator (DAX) even if it is not fully serverless. In addition, AWS ElastiCache provides Redis and Memcached, but has not yet provided them as fully managed services. In addition, Google App Engine provides Memcache as an in-memory caching service, but it can only be accessed in App Engine itself.

Conclusion

With the growth and high adaptation of serverless computing in the IT industry, serverless storage is growing rapidly. We hope that these services will be continuously optimized and improved to help developers achieve true serverless application development.

Original source: Medium author: Rajind Ruparathna

Original link: https://medium.com/@rajindruparathna/what-makes-a-storage-service-truly-serverless-93987826039e

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.