#alloydb — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #alloydb, aggregated by home.social.
-
Почему мы выбрали рекурсивные SQL-запросы вместо GraphQL для графа знаний
Сравниваем нативный Property Graph в Spanner с рекурсивными CTE в AlloyDB — и объясняем, почему для персональной wiki второй подход оказался практичнее
-
Как мы с моим ботом OpenClaw сделали ему семантическую память на AlloyDB Omni за полчаса
История о том, как превратить ИИ-агента из "золотой рыбки" с памятью в пределах одной сессии в полноценного цифрового сотрудника с графовым хранилищем знаний.
-
Семантический обновляемый кэш на AlloyDB Omni
Предположим, вы построили RAG-сервис на SQL, и он отлично работает. Довольно быстро, очень точно, и очень дорого , ведь каждый запрос к сервису требует обращения к LLM для генерации ответа по чанкам, извлеченным из базы знаний. И чем больше мы извлекли таких фрагментов, тем больше входных токенов тратится на составной промпт, даже если ответ будет состоять из одного предложения. Можно, конечно, заранее срезать количество извлекаемых чанков, но это отразится на качестве ответов. Можно настроить кэш, который экономит на обращениях к сервису, когда приходят одинаковые вопросы. Но когда пользователь спрашивает "How to get developer support?”, и тут же другой пользователь спрашивает "How to ask development-related questions?", ваш сервис каждый раз будет генерировать ответ заново, сжигая ваши токены и заставляя пользователя ждать. Обычный кэш тут бессилен: для него эти две фразы — абсолютно разные ключи. В этой статье я расскажу, как развернуть мощный семантический кэш на базе AlloyDB Omni (PostgreSQL от Google), используя векторный поиск ScaNN, автоматическое партиционирование и планировщик задач. Мы пройдём путь от настройки Docker-контейнера до продакшн-архитектуры.
-
Postgres по-русски: где наши Aurora, AlloyDB и Neon?
Базы данных давно являются фундаментом цифровой экономики. От их архитектуры и производительности во многом зависят скорость вывода продуктов на рынок, стабильность сервисов и итоговая стоимость ИТ-инфраструктуры. В мировой практике одним из основных стандартов де-факто, вокруг которого формируются экосистемы серьезных решений, стала открытая СУБД PostgreSQL. В России она используется во множестве корпоративных приложений, есть целый ряд отечественных форков и дистрибутивов. Но у ряда зарубежных компаний( есть серьезные прорывные реализации, интенсивно развивающие Postgres (например, Aurora, AlloyDB и Neon, об этом ниже), а у российских этого почему-то не наблюдается. Это противоречие между массовым использованием PostgreSQL в нашей стране и отсутствием технологического прорыва задает остроту сегодняшней повестки отечественного СУБД-строения.
https://habr.com/ru/companies/tantor/articles/988340/
#neon #horizondb #aurora #alloydb #alibaba #postgresql #tantor_xdata #xdata #tantor_postgres
-
All benchmarks are liars.
Also, here’s a set of db benchmarks from #clickhouse:
https://benchmark.clickhouse.com/
Various DBs including:
- #mysql
- #postgres
- #sqlite
- #timescaledb
- #alloydb
- #mariadb -
All benchmarks are liars.
Also, here’s a set of db benchmarks from #clickhouse:
https://benchmark.clickhouse.com/
Various DBs including:
- #mysql
- #postgres
- #sqlite
- #timescaledb
- #alloydb
- #mariadb -
🆕 Google introduced a free trial for #AlloyDB - their fully managed PostgreSQL-compatible database service.
The trial allows users to test AlloyDB's capabilities with their workloads for up to 30 days.
More insights available on #InfoQ👉 https://bit.ly/4ckHoml
-
🆕 Google introduced a free trial for #AlloyDB - their fully managed PostgreSQL-compatible database service.
The trial allows users to test AlloyDB's capabilities with their workloads for up to 30 days.
More insights available on #InfoQ👉 https://bit.ly/4ckHoml
-
New episode: "Decoupled storage and compute"
Nikolay and Michael discuss a listener question — about products that take #Postgres and transform it to something that decouples compute from storage (RDS #Aurora, GC #AlloyDB, #Neon etc.) and whether they see something like this landing upstream in the medium term.
🎙️ https://postgres.fm/episodes/decoupled-storage-and-compute
-
New episode: "Decoupled storage and compute"
Nikolay and Michael discuss a listener question — about products that take #Postgres and transform it to something that decouples compute from storage (RDS #Aurora, GC #AlloyDB, #Neon etc.) and whether they see something like this landing upstream in the medium term.
🎙️ https://postgres.fm/episodes/decoupled-storage-and-compute
-
#Performance of your #database is a huge topic. One critical piece is running benchmarks, but do you really know how/why/when to benchmark?
Check out this blog post by @sky_vst on the philosophy of benchmarking #AlloyDB!
-
#Performance of your #database is a huge topic. One critical piece is running benchmarks, but do you really know how/why/when to benchmark?
Check out this blog post by @sky_vst on the philosophy of benchmarking #AlloyDB!
-
#Google Announces Preview of #AlloyDB Omni: Run a #PostgreSQL-Compatible Database Anywhere https://www.infoq.com/news/2023/04/google-cloud-alloydb-omni/
-
#Google Announces Preview of #AlloyDB Omni: Run a #PostgreSQL-Compatible Database Anywhere https://www.infoq.com/news/2023/04/google-cloud-alloydb-omni/
-
Getting ready for the first #opensource data infrastructure meetup in Milan!
Let’s talk about #database with #clickhouse and #alloydb!
Good food and good tech!
@aiven_io
-
Getting ready for the first #opensource data infrastructure meetup in Milan!
Let’s talk about #database with #clickhouse and #alloydb!
Good food and good tech!
@aiven_io
-
Happening tonight!
Join us in Milan at 6PM for a night concentrated on #database: #AlloyDB from @googlecloud and #ClickHouse!
-
Happening tonight!
Join us in Milan at 6PM for a night concentrated on #database: #AlloyDB from @googlecloud and #ClickHouse!
-
Getting your indexes right on #PostgreSQL can be an ever-evolving game of whack-a-mole with engineering. #AlloyDB builds in AI to help advise you on what indexes to create based on a multitude of factors. Check out this blog to explain it all in depth:
-
Getting your indexes right on #PostgreSQL can be an ever-evolving game of whack-a-mole with engineering. #AlloyDB builds in AI to help advise you on what indexes to create based on a multitude of factors. Check out this blog to explain it all in depth:
-
If you're around Milan and into the #Database world, this event is a must!
Milan Open Source Data Infrastructure Meetup on the 12th April to hear about two of the new cool databases: #AlloyDB and #ClickHouse from @googlecloud and @aiven_io
https://www.meetup.com/milano-open-source-data-infrastructure-meetup/events/292063757/
-
If you're around Milan and into the #Database world, this event is a must!
Milan Open Source Data Infrastructure Meetup on the 12th April to hear about two of the new cool databases: #AlloyDB and #ClickHouse from @googlecloud and @aiven_io
https://www.meetup.com/milano-open-source-data-infrastructure-meetup/events/292063757/
-
You've all probably heard me say #AlloyDB is #PostgreSQL compatible wha, a thousand times (if not maybe I'm not trying hard enough)? Well, it is, and to hammer the point home even more, we made an awesome animated video about it too:
https://www.youtube.com/watch?v=88k-eB3Eon8&list=PLIivdWyY5sqLCsehDPydpIZ31Rmrm0TI9
-
You've all probably heard me say #AlloyDB is #PostgreSQL compatible wha, a thousand times (if not maybe I'm not trying hard enough)? Well, it is, and to hammer the point home even more, we made an awesome animated video about it too:
https://www.youtube.com/watch?v=88k-eB3Eon8&list=PLIivdWyY5sqLCsehDPydpIZ31Rmrm0TI9
-
I think it's fair to say that when #Google announced #AlloyDB, there was a certain amount of scepticism, given how strong #AWS #Aurora is in the market. However, the uptake seems to be pretty good. Check out all these certified vendors (including #SnapLogic): https://www.datanami.com/2023/03/29/partners-line-up-for-google-cloud-ready-for-alloydb-designation/ #database #DBaaS
-
I think it's fair to say that when #Google announced #AlloyDB, there was a certain amount of scepticism, given how strong #AWS #Aurora is in the market. However, the uptake seems to be pretty good. Check out all these certified vendors (including #SnapLogic): https://www.datanami.com/2023/03/29/partners-line-up-for-google-cloud-ready-for-alloydb-designation/ #database #DBaaS
-
📣 #SnapLogic is now in the #GoogleCloud Marketplace: https://www.snaplogic.com/company/newsroom/press-releases/snaplogic-delivers-automated-data-driven-insights-through-google-cloud-marketplace And that's not all! We also achieved the #Google #Cloud Ready – #AlloyDB designation. A big endorsement, and a way to make it even easier to get access to the power of the SnapLogic platform? Yes please! #iPaaS #LowCode #NoCode #DataIntegration
-
📣 #SnapLogic is now in the #GoogleCloud Marketplace: https://www.snaplogic.com/company/newsroom/press-releases/snaplogic-delivers-automated-data-driven-insights-through-google-cloud-marketplace And that's not all! We also achieved the #Google #Cloud Ready – #AlloyDB designation. A big endorsement, and a way to make it even easier to get access to the power of the SnapLogic platform? Yes please! #iPaaS #LowCode #NoCode #DataIntegration
-
[ Google Cloud’s AlloyDB Omni lets you run its PostgreSQL-compatible database anywhere | TechCrunch ]
https://techcrunch.com/2023/03/29/google-cloud-launches-alloydb-omni-a-downloadable-version-of-its-postgresql-compatible-database/ #AlloyDB #PostgreSQL #Google -
My product team at Google just released a downloadable edition of #AlloyDB, its #PostgreSQL-compatible database engine loaded with lots of neat features like an index advisor and a columnar engine and so on and suchlike.
It's in preview! You can install it on any newish RHEL or Debian-like and try it out for free.
https://cloud.google.com/blog/products/databases/run-alloydb-anywhere
I co-authored the docs, by gar.
-
My product team at Google just released a downloadable edition of #AlloyDB, its #PostgreSQL-compatible database engine loaded with lots of neat features like an index advisor and a columnar engine and so on and suchlike.
It's in preview! You can install it on any newish RHEL or Debian-like and try it out for free.
https://cloud.google.com/blog/products/databases/run-alloydb-anywhere
I co-authored the docs, by gar.
-
Not always my fault if my blog titles are too long. I've just added "Continuous Backup" as that's how #AlloyDB markets it 😂
---
RT @FranckPachot
What is your @PostgreSQL alternative to 🅾 Oracle Flashback Database / 🔶 Amazon Aurora Backtrack?
Is transactional DDL sufficient to protect application releases with DDL + DML on the DB?
🚀 @Yugabyte has distributed snapshots:
https://dev.to/yugabyte/pitr-snapshot-an-easy-flashback-backtrack-f…
https://twitter.com/FranckPachot/status/1536808557974327298 -
Not always my fault if my blog titles are too long. I've just added "Continuous Backup" as that's how #AlloyDB markets it 😂
---
RT @FranckPachot
What is your @PostgreSQL alternative to 🅾 Oracle Flashback Database / 🔶 Amazon Aurora Backtrack?
Is transactional DDL sufficient to protect application releases with DDL + DML on the DB?
🚀 @Yugabyte has distributed snapshots:
https://dev.to/yugabyte/pitr-snapshot-an-easy-flashback-backtrack-f…
https://twitter.com/FranckPachot/status/1536808557974327298 -
Hoooo boy folks, it's here! You may have heard me talking a lot about #AlloyDB. Well now it's officially GA! Check out the launch blog, try it out, and see what super-powered #postgresql can do!
Can also about our partner ecosystem and the first salvo of products that integrate with AlloyDB:
-
Hoooo boy folks, it's here! You may have heard me talking a lot about #AlloyDB. Well now it's officially GA! Check out the launch blog, try it out, and see what super-powered #postgresql can do!
Can also about our partner ecosystem and the first salvo of products that integrate with AlloyDB:
-
@richard4444 I wanted to call #AlloyDB ++Postgres but they shut me down. If you had a magic wand, what’s a key message we should be spamming out there about Spanner to get more of the word out? Also possible we aren’t reaching the right audiences (we’ve put out a ton of content for folks to get familiar with it but mostly on our Cloud Blog). Hard to give conference talks on it because no one likes a vendor talk, and sponsored talks are back in the echo chamber for the most part.
-
@garystafford I’m admittedly not as familiar with Aurora (because Google) but I will say that with an architecture of tons of clients, be it microservices, or mobile users, or web users, etc, you need to scale the infra save as front end. Some kind of load balancing is a must. Google’s new database offering, #AlloyDB has built in read replica pools with load balancing built into the pool itself. So scaling the read side is really simple.