---
title: "Comment by Jorge Romero on \"How We Leverage Serverless for Backing up Your Posts.\" | Hashnode"
description: "That was really insightful! I learned a lot. Thank you Sandro Volpicella I wasn't able to follow all of the DynamoDB part. Mainly because I have never used"
url: "https://hashnode.com/posts/serverless-github-backups/62443f47fca8a5eb4a9e6594/comment/62449583114d74ee1e63219d"
---

[Discussion](https://hashnode.com/posts/serverless-github-backups/62443f47fca8a5eb4a9e6594)

Comment [How We Leverage Serverless for Backing up Your Posts.](https://hashnode.com/posts/serverless-github-backups/62443f47fca8a5eb4a9e6594)

[Jorge Romero](https://hashnode.com/@JorgeRL)

[Mar 30, 2022](https://hashnode.com/posts/serverless-github-backups/62443f47fca8a5eb4a9e6594/comment/62449583114d74ee1e63219d)

That was really insightful! I learned a lot. Thank you [Sandro Volpicella](https://hashnode.com/@SandroVolpicella)

-   I wasn't able to follow all of the DynamoDB part. Mainly because I have never used it. But I got the broad idea.
    
-   It seems like a lot more work went into getting that in-progress feedback than the actual backup feature. I look forward to read how you iterate on that!
    
-   I also have a small question. `publicationID` belongs to a user or to a blog? I am guessing it is for each blog. So any user can have several publicationIDs? I am not familiar with Hashnode's API, but I have been toying around with it a little. And it seems to me that a user can only have a single publication (???) 😅
    

I am just figuring serverless out so this helps a lot!

By the way, I love the drawings :D

[Sandro Volpicella](https://hashnode.com/@SandroVolpicella)

[Mar 31, 2022](https://hashnode.com/posts/serverless-github-backups/62443f47fca8a5eb4a9e6594/comment/62449583114d74ee1e63219d#reply-6245383c93b269f2e37423b7)

Thanks for your comment!

I get that the whole approach is a bit confusing or seems a lot for a simple feature like that. But this shows some complexities of an async process.

**DynamoDB**: DynamoDB is mainly there for getting the connection back to the user. Since we don't have a websocket server such as socket.io we need to store the `connectionId` somewhere to respond to the user. This is what DynamoDB is doing.

**Feedback** Yes a huge part was definitely the whole feedback loop for the enduser. This infrastructure builds the basics now because we can reuse it for other features as well.

**PublicationID** A `publicationId` belongs to a publication. Internally we distinguish between a user and a publication. A user can have several publications, e.g. one team publication and one private publication.

I hope that helps. Drawings are in excalidraw which is an awesome tool! 🙂