I have a question about the response time when accessing through HTTP such as AWS DynamoDB and Elasticsearch.
No matter how fast DynamoDB and Elasticsearch are, it will take about 100 ms more to access those databases through HTTP when you access them through applications (for example, Rails).
Is this inevitable?I wonder why it is designed like this.
If I could access it like MySQL, I could cut the response time for HTTP requests, so it would be faster, but I was wondering why it was via HTTP.
Thank you for your cooperation.
aws elasticsearch
Is this inevitable?Why is this specification?
This service is intended to be used directly via the Internet from mobile apps and client-side applications (such as SPAs), so I think it is via HTTP.
If you go over the Internet, there are proxies and firewalls on the way, so HTTP is used without being blocked.
If the service is intended to be used as the backend of the web server, the communication protocol that you can choose has a degree of freedom, so I think it will focus on response time and socket communication.
Why is this specification?
Because products such as Elasticsearch adopted REST, I think the communication protocol became HTTP due to that restriction.
For example, if you connect like Mysql, you will need to manage the state of the session and it will no longer be REST.
Why REST? is a generalization, but I think it's because I wanted to make it expandable and simple to use by making it a RESTful system.
© 2024 OneMinuteCode. All rights reserved.