---
title: "Comment by Daniel Ziltener on \"Database Architectures & Use Cases - Explained\" | Hashnode"
description: "Time series databases don&#39;t necessarily suffer in write speeds. A good example for that is Datomic, it can easily parallelize reads almost arbitrarily. And"
url: "https://hashnode.com/posts/database-architectures-and-use-cases-explained/5f60f74ac417c20b18b72286/comment/5f622983456ff01582e81184"
---

[Discussion](https://hashnode.com/posts/database-architectures-and-use-cases-explained/5f60f74ac417c20b18b72286)

Comment [Database Architectures & Use Cases - Explained](https://hashnode.com/posts/database-architectures-and-use-cases-explained/5f60f74ac417c20b18b72286)

[Daniel Ziltener](https://hashnode.com/@zilti)

[Sep 16, 2020](https://hashnode.com/posts/database-architectures-and-use-cases-explained/5f60f74ac417c20b18b72286/comment/5f622983456ff01582e81184)

Time series databases don't necessarily suffer in write speeds. A good example for that is Datomic, it can easily parallelize reads almost arbitrarily. And that is possible exactly because the database is append-only (and transactions are also supported). Since it is append-only, the whole database can be seen as an atomic value, and thus no locking or anything has to be done.

[Margo McCabe](https://hashnode.com/@MargoHDB)

[Sep 16, 2020](https://hashnode.com/posts/database-architectures-and-use-cases-explained/5f60f74ac417c20b18b72286/comment/5f622983456ff01582e81184#reply-5f62302d018bef1561b429ed)

Thank you for your feedback! In our teams experience, we've seen that time series databases are incredibly fast at writing data, because, like you said, everything is linear. We have found that read speeds for specific queries tend to suffer relative to write speeds. Our team has not worked with Datomic, so that may very well be the exception in this case.
