[README]:consistencyModel #346
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
vaibhav5140
temporarily deployed
to
integration-tests
GitHub Actions
Inactive
fuatbasik
requested changes
Aug 18, 2025
|
|
||
| ## Consistency Model | ||
|
|
||
| Analytics Accelerator Library for Amazon S3 implements a time-based consistency model using TTL (Time-To-Live) for metadata caching. This ensures bounded staleness while maintaining performance. The library caches object metadata with a configurable TTL period. When metadata TTL expires, the library performs a conditional HEAD request to S3 to verify if the object has changed. If the object is unchanged, the TTL period is reset. If the object has changed, the cache is updated with new metadata. |
There was a problem hiding this comment.
we can drop conditional from the sentence.
|
|
||
| Within-Stream Consistency: Each stream maintains strict consistency throughout its lifetime by locking to a specific object version. If cached blocks are evicted and the object has changed in S3, subsequent reads will fail with 412 Precondition Failed errors rather than serving inconsistent data. | ||
|
|
||
| Cross-Stream Consistency: Streams created within the same TTL window see the same object version. When an S3 object is modified, existing streams continue using their original version, while new streams will see the updated version after TTL expiration. This can result in different streams temporarily accessing different versions of the same object, but the maximum staleness is bounded by the configured TTL duration. |
There was a problem hiding this comment.
I think you can drop the sentence This can result ..... It might confuse to reader.
|
|
||
| Cross-Stream Consistency: Streams created within the same TTL window see the same object version. When an S3 object is modified, existing streams continue using their original version, while new streams will see the updated version after TTL expiration. This can result in different streams temporarily accessing different versions of the same object, but the maximum staleness is bounded by the configured TTL duration. | ||
|
|
||
| The metadata cache may evict entries when reaching its size limit (5000 entries by default), triggering fresh metadata fetching on the next access without affecting existing streams consistency. |
There was a problem hiding this comment.
We might want to add a sentence to describe the reasoning behind having a size-limit in addition. Something like, To avoid an ever-growing metdata cache, when TTL is set to a very large number, we also use a threshold to limit number of metadata entries kept in memory.. Next, you should also tell what is the eviction strategy when this threshold is reached.
| * `metadatastore.capacity` - Maximum size for metadata cache entries, default is 5000 | ||
|
|
||
| TTL = 0 (Strong Consistency): | ||
| Setting `metadata.ttl.default=0` provides the strongest consistency guarantees. Every metadata access triggers a fresh HEAD request to S3. This configuration results in higher latency due to increased S3 requests but is recommended when absolute freshness is required. |
There was a problem hiding this comment.
i think saying provides Strong Consistency is enough. Since Strong Consistency is a technical term, i am not sure if we can use stronger here.
| TTL = 0 (Strong Consistency): | ||
| Setting `metadata.ttl.default=0` provides the strongest consistency guarantees. Every metadata access triggers a fresh HEAD request to S3. This configuration results in higher latency due to increased S3 requests but is recommended when absolute freshness is required. | ||
|
|
||
| High TTL Values (Relaxed Consistency): |
There was a problem hiding this comment.
you mean eventual consistency. We are not talking about Relaxed consistency models here.
| Setting `metadata.ttl.default=0` provides the strongest consistency guarantees. Every metadata access triggers a fresh HEAD request to S3. This configuration results in higher latency due to increased S3 requests but is recommended when absolute freshness is required. | ||
|
|
||
| High TTL Values (Relaxed Consistency): | ||
| Setting a high TTL value (e.g., `metadata.ttl.default` for 1 hour) allows for longer caching of metadata. This configuration offers better performance due to reduced S3 requests and is suitable for static data or when staleness can be tolerated. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description of change
[README Update]
Follow-Up PR of (#338) to include Consistency Model documentation of AAL
Does this contribution need a changelog entry?
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license and I agree to the terms of the Developer Certificate of Origin (DCO).