[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.] #  @debasishg Debasish (দেবাশিস্) Ghosh 🇮🇳 Debasish (দেবাশিস্) Ghosh 🇮🇳 posts on X about skip, vertical, filter, bloom the most. They currently have XXXXXX followers and XXX posts still getting attention that total XXXXX engagements in the last XX hours. ### Engagements: XXXXX [#](/creator/twitter::6562002/interactions)  - X Week XXXXXX +54% - X Month XXXXXX +98% - X Months XXXXXXX -XX% - X Year XXXXXXX -XXXX% ### Mentions: XX [#](/creator/twitter::6562002/posts_active)  - X Week XX +11% - X Month XX +43% - X Months XXX -XX% - X Year XXX +99% ### Followers: XXXXXX [#](/creator/twitter::6562002/followers)  - X Week XXXXXX +0.82% - X Month XXXXXX +2.80% - X Months XXXXXX +4.10% - X Year XXXXXX +15% ### CreatorRank: XXXXXXX [#](/creator/twitter::6562002/influencer_rank)  ### Social Influence [#](/creator/twitter::6562002/influence) --- **Social category influence** [stocks](/list/stocks) [technology brands](/list/technology-brands) **Social topic influence** [skip](/topic/skip) #278, [vertical](/topic/vertical), [filter](/topic/filter), [bloom](/topic/bloom), [supported](/topic/supported), [async](/topic/async), [over the](/topic/over-the), [als](/topic/als), [events](/topic/events), [peek](/topic/peek) ### Top Social Posts [#](/creator/twitter::6562002/posts) --- Top posts by engagements in the last XX hours "@TheEduardoRFS May be just the right time to get hold of" [X Link](https://x.com/debasishg/status/1978513428554641424) [@debasishg](/creator/x/debasishg) 2025-10-15T17:28Z 12.1K followers, XXX engagements "@curlykoder If u r interested take a look at @andy_pavlos Advanced Database Systems course at @CMUDB where he covers Umbra in great details" [X Link](https://x.com/debasishg/status/1979030505321365754) [@debasishg](/creator/x/debasishg) 2025-10-17T03:43Z 12.1K followers, XX engagements "Skip lists are usually known not to be too cache friendly. But spent quite some time digging into the source code to understand how Redis implements skip lists with great cache locality. BTW Sorted Sets in Redis are implemented with skip lists. The design choices that Redis makes mean the skip list pays a small pointer-chasing cost for the O(log N) seek but enjoys cache-friendly linear access for the heavy lifting (ranges/iterations). For tiny zsets Redis sidesteps the issue entirely with a contiguous encoding. Had a great code reading session for the above understanding " [X Link](https://x.com/debasishg/status/1979929066741584038) [@debasishg](/creator/x/debasishg) 2025-10-19T15:14Z 12.1K followers, 8397 engagements "Exactly thats what Redis did with skip list. Ensuring sequential access keeping cache lines hot reducing allocations are all implementation characteristics that the definition of the data structure doesnt have. So my suggestion would be to go for the simplest data structure that works. As in the case of Redis they could have done with trees. But the implementation would have been complex and balancing would have thrown elements out of cache lines. You can study the current skip list data structure in Redis source code for this" [X Link](https://x.com/debasishg/status/1980111862554612027) [@debasishg](/creator/x/debasishg) 2025-10-20T03:20Z 12.1K followers, XX engagements "New paper on LSM tree dropped in SIGMOD-25. It gives new insights into the trade-offs in vertical (adopted by RocksDB LevelDB Cassandra) and horizontal growth (used by AsterixDB BigTable HBase) of an LSM tree as it ingests new data. It suggests a technique that offers better write performance than the vertical scheme while incurring about six times less additional space cost compared to the horizontal scheme. link: 👇" [X Link](https://x.com/debasishg/status/1929052744796811474) [@debasishg](/creator/x/debasishg) 2025-06-01T05:49Z 12.1K followers, 6774 engagements "Sometimes you need to read the comments in a production grade implementation to feel the engineering aspects of a data structure or an algorithm. No book will give you this lesson . 👇from bloom filter implementation in RocksDB " [X Link](https://x.com/debasishg/status/1952774668928057795) [@debasishg](/creator/x/debasishg) 2025-08-05T16:52Z 12.1K followers, 7544 engagements "Which of the popular LSM based storage engines take advantage of io_uring on Linux I was doing some fact finding. - Looks like RocksDB w/ POSIX env can use io_uring for parallel async reads (notably MultiGet) and broader async-I/O paths. This is supported on Linux = XXX. - ScyllaDB via Seastar the I/O framework ScyllaDB uses has an io_uring reactor backend - Pebble/CockroachDB are working on it particularly for parallelizing lookups as the io_uring system is used in RocksDB's MultiGet for parallelism rather than threads - Maybe TiKV doesnt ship io_uring in production but there have been some" [X Link](https://x.com/debasishg/status/1969093306710675476) [@debasishg](/creator/x/debasishg) 2025-09-19T17:36Z 12.1K followers, 2449 engagements "Here's Scala X . extension A(x: A) infix def B(f: A = B): B = f(x)" [X Link](https://x.com/debasishg/status/1975638602865647800) [@debasishg](/creator/x/debasishg) 2025-10-07T19:05Z 12.1K followers, 4362 engagements "LeanStore was a landmark in in-memory data management beyond main memory and had influenced lots of subsequent implementations like Umbra and MotherDuck (DuckDB in the cloud). Yet it all started with Graefe et al's paper that proposed using lazy pointer swizzling for relational buffer managers. Here's a rough chronological list of how buffer management evolved over the years and is still being researched for improvements like VMCache trying to eliminate swizzling costs through virtual memory integration in production quality storage engines. Looks like I will be quite busy munching on these" [X Link](https://x.com/debasishg/status/1978878968481681456) [@debasishg](/creator/x/debasishg) 2025-10-16T17:41Z 12.1K followers, 3067 engagements "@aspender I think the LI post may not be visible to everybody . here's a snapshot " [X Link](https://x.com/debasishg/status/1979565939533467789) [@debasishg](/creator/x/debasishg) 2025-10-18T15:11Z 12.1K followers, XX engagements "A podcast that gives you a peek into the internals of Monocle a high performance time series storage engine and Husky an event store that can store and query trillions of events per day - both from DataDog and both implemented in Rust" [X Link](https://x.com/debasishg/status/1979833867910754482) [@debasishg](/creator/x/debasishg) 2025-10-19T08:55Z 12.1K followers, 10.8K engagements "ICYMI this is just a PoC where I am trying to validate an architectural style "DDD-at-the-edges with DoD-as-an-implementation" which should give a good balance between domain friendly APIs for the developers while ensuring data oriented design performance at the backend. Much of the documentation and code are generated by Claude and might have some scope for optimization " [X Link](https://x.com/debasishg/status/1980169918856257815) [@debasishg](/creator/x/debasishg) 2025-10-20T07:11Z 12.1K followers, XXX engagements
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]
@debasishg Debasish (দেবাশিস্) Ghosh 🇮🇳
Debasish (দেবাশিস্) Ghosh 🇮🇳 posts on X about skip, vertical, filter, bloom the most. They currently have XXXXXX followers and XXX posts still getting attention that total XXXXX engagements in the last XX hours.
Engagements: XXXXX #
- X Week XXXXXX +54%
- X Month XXXXXX +98%
- X Months XXXXXXX -XX%
- X Year XXXXXXX -XXXX%
Mentions: XX #
- X Week XX +11%
- X Month XX +43%
- X Months XXX -XX%
- X Year XXX +99%
Followers: XXXXXX #
- X Week XXXXXX +0.82%
- X Month XXXXXX +2.80%
- X Months XXXXXX +4.10%
- X Year XXXXXX +15%
CreatorRank: XXXXXXX #
Social Influence #
Social category influence stocks technology brands
Social topic influence skip #278, vertical, filter, bloom, supported, async, over the, als, events, peek
Top Social Posts #
Top posts by engagements in the last XX hours
"@TheEduardoRFS May be just the right time to get hold of"
X Link @debasishg 2025-10-15T17:28Z 12.1K followers, XXX engagements
"@curlykoder If u r interested take a look at @andy_pavlos Advanced Database Systems course at @CMUDB where he covers Umbra in great details"
X Link @debasishg 2025-10-17T03:43Z 12.1K followers, XX engagements
"Skip lists are usually known not to be too cache friendly. But spent quite some time digging into the source code to understand how Redis implements skip lists with great cache locality. BTW Sorted Sets in Redis are implemented with skip lists. The design choices that Redis makes mean the skip list pays a small pointer-chasing cost for the O(log N) seek but enjoys cache-friendly linear access for the heavy lifting (ranges/iterations). For tiny zsets Redis sidesteps the issue entirely with a contiguous encoding. Had a great code reading session for the above understanding "
X Link @debasishg 2025-10-19T15:14Z 12.1K followers, 8397 engagements
"Exactly thats what Redis did with skip list. Ensuring sequential access keeping cache lines hot reducing allocations are all implementation characteristics that the definition of the data structure doesnt have. So my suggestion would be to go for the simplest data structure that works. As in the case of Redis they could have done with trees. But the implementation would have been complex and balancing would have thrown elements out of cache lines. You can study the current skip list data structure in Redis source code for this"
X Link @debasishg 2025-10-20T03:20Z 12.1K followers, XX engagements
"New paper on LSM tree dropped in SIGMOD-25. It gives new insights into the trade-offs in vertical (adopted by RocksDB LevelDB Cassandra) and horizontal growth (used by AsterixDB BigTable HBase) of an LSM tree as it ingests new data. It suggests a technique that offers better write performance than the vertical scheme while incurring about six times less additional space cost compared to the horizontal scheme. link: 👇"
X Link @debasishg 2025-06-01T05:49Z 12.1K followers, 6774 engagements
"Sometimes you need to read the comments in a production grade implementation to feel the engineering aspects of a data structure or an algorithm. No book will give you this lesson . 👇from bloom filter implementation in RocksDB "
X Link @debasishg 2025-08-05T16:52Z 12.1K followers, 7544 engagements
"Which of the popular LSM based storage engines take advantage of io_uring on Linux I was doing some fact finding. - Looks like RocksDB w/ POSIX env can use io_uring for parallel async reads (notably MultiGet) and broader async-I/O paths. This is supported on Linux = XXX. - ScyllaDB via Seastar the I/O framework ScyllaDB uses has an io_uring reactor backend - Pebble/CockroachDB are working on it particularly for parallelizing lookups as the io_uring system is used in RocksDB's MultiGet for parallelism rather than threads - Maybe TiKV doesnt ship io_uring in production but there have been some"
X Link @debasishg 2025-09-19T17:36Z 12.1K followers, 2449 engagements
"Here's Scala X . extension A(x: A) infix def B(f: A = B): B = f(x)"
X Link @debasishg 2025-10-07T19:05Z 12.1K followers, 4362 engagements
"LeanStore was a landmark in in-memory data management beyond main memory and had influenced lots of subsequent implementations like Umbra and MotherDuck (DuckDB in the cloud). Yet it all started with Graefe et al's paper that proposed using lazy pointer swizzling for relational buffer managers. Here's a rough chronological list of how buffer management evolved over the years and is still being researched for improvements like VMCache trying to eliminate swizzling costs through virtual memory integration in production quality storage engines. Looks like I will be quite busy munching on these"
X Link @debasishg 2025-10-16T17:41Z 12.1K followers, 3067 engagements
"@aspender I think the LI post may not be visible to everybody . here's a snapshot "
X Link @debasishg 2025-10-18T15:11Z 12.1K followers, XX engagements
"A podcast that gives you a peek into the internals of Monocle a high performance time series storage engine and Husky an event store that can store and query trillions of events per day - both from DataDog and both implemented in Rust"
X Link @debasishg 2025-10-19T08:55Z 12.1K followers, 10.8K engagements
"ICYMI this is just a PoC where I am trying to validate an architectural style "DDD-at-the-edges with DoD-as-an-implementation" which should give a good balance between domain friendly APIs for the developers while ensuring data oriented design performance at the backend. Much of the documentation and code are generated by Claude and might have some scope for optimization "
X Link @debasishg 2025-10-20T07:11Z 12.1K followers, XXX engagements