[freedomben]

Sometimes yes, I've seen it. It even tends to happen on NoSQL databases as well. Three times I've seen apps start on top of Dynamo DB, and then end up re-implementing relational databases at the application level anyway. Starting with postgres would have been the right answer for all three of those. Initial dev went faster, but tech debt and complexity quickly started soaking up all those gains and left a hard-to-maintain mess.

[leafarlua]

This always confuses me because we have decades of SQL and all its issues as well. Hundreds of experienced devs talking about all the issues in SQL and the quirks of queries when your data is not trivial.

One would think that for a startup of sorts, where things changes fast and are unpredictable, NoSQL is the correct answer. And when things are stable and the shape of entities are known, going for SQL becomes a natural path.

There is also cases for having both, and there is cases for graph-oriented databases or even columnar-oriented ones such as duckdb.

Seems to me, with my very limited experience of course, everything leads to same boring fundamental issue: Rarely the issue lays on infrastructure, and is mostly bad design decisions and poor domain knowledge. Realistic, how many times the bottleneck is indeed the type of database versus the quality of the code and the imlementation of the system design?

[marcosdumay]

No, when things change fast and unpredictably, NoSQL is worse than when they are well-known and stable.

NoSQL gains you no speed at all in redesigning your system. Instead, you trade a few hard to do tasks in data migration into an unsurmountable mess of data inconsistency bugs that you'll never actually get into the end of.

> is mostly bad design decisions and poor domain knowledge

Yes, using NoSQL to avoid data migrations is a bad design decision. Usually created by poor general knowledge.

[james_marks]

If the argument for NoSQL is, “we don’t know what our schema is going to be”, stop.

Stop and go ask more questions until you have a better understanding of the problem.

[jampekka]

Oftentimes better understanding of the problem needs trying out solutions. Armchair architectures tend to blow up in contact with reality.

[freedomben]

For sure, though with databases it's usually pretty clear even at the start whether your "objects" will be relational in nature. I can't think of a single time that hasn't been the case, over hundreds of apps/services I've been part of. Things like asynchronous jobs, message queues, even object storage, I fully agree though.

[ranger_danger]

Even a JSON column would be better in most cases IMO, and on Postgres you can also make indexes on json keys.

[leafarlua]

Makes sense. But in this case, why NoSQL exists? What problems does it resolves and when should it be considered? I'm being naive, but fast changing environment has been one of the main advantages that I was taught from devs when it comes to NoSQL vs SQL (nosql being the choice for flexible schemas). So it is more about BASE vs ACID?

[marcosdumay]

NoSQL was created to deal with scales where ACID becomes a bottleneck. It also shown itself useful for dealing with data that don't actually have an schema.

If you have either of those problems, you will know it very clearly.

Also, ironically, Postgres became one of the most scalable NoSQL bases out there, and one of the most flexible to use unstructured data too.

[freedomben]

Agreed. In my experience (YMMV), there was also a real adoption push in the js world from primarily front-end people that wanted to do some backend but didn't want to learn/deal with SQL databases. I don't say that with malice, I was also on-board the NoSQL train for a bit before I actually gained experience with the headaches it caused. The appeal of "just dump your JSON blob straight in" was (and still is) strong. Software is all about learning, and sometimes that learning is expensive. We've all built something we later regretted.

[yfontana]

As a data architect I dislike the term NoSQL and often recommend that my coworkers not use it in technical discussions, as it is too vague. Document, key-value and graph DBs are usually considered NoSQL, but they have fairly different use cases (and I'd argue that search DBs like Elastic / OpenSearch are in their own category as well).

To me write scaling is the main current advantage of KV and document DBs. They can generally do schema evolution fairly easily, but nowadays so can many SQL DBs, with semi-structured column types. Also, you need to keep in mind that KV and document DBs are (mostly) non-relational. The more relational your data, the less likely you are to actually benefit from using those DBs over a relational, SQL DB.

[gf000]

Probably the best use case would be something like a Facebook profile page for a given user.

It may not have a very rigid schema, you may later add several other optional fields.

You need very large scale (as in no of concurrent accesses), you want to shard the data by e.g. location. But also, the data is not "critical", your highschool not being visible temporarily for certain users is not an issue.

You mostly use the whole dataset "at the same time", you don't do a lot of WHERE, JOIN on some nested value.

In every other case I would rather reach for postgres with a JSONB column.

[ngc248]

NOSQL works out really well when you know the access patterns you are working with, if not it will be like pegging a round hole with a square peg

[tracker1]

I think part of it is the scale in terms of the past decade and a half... The hardware and vertical scale you could get in 2010 is dramatically different than today.

A lot of the bespoke no-sql data stores really started to come to the forefront around 2010 or so. At that time, having 8 cpu cores and 10k rpm SAS spinning drives was a high end server. Today, we have well over 100 cores, with TBs of RAM and PCIe Gen 4/5 NVME storage (u.x) that is thousands of times faster and has a total cost lower than the servers from 2010 or so that your average laptop can outclass today.

You can vertically scale a traditional RDBMS like PostgreSQL to an extreme degree... Not to mention utilizing features like JSONB where you can have denormalized tables within a structured world. This makes it even harder to really justify using NoSQL/NewSQL databases. The main bottlenecks are easier to overcome if you relax normalization where necessary.

There's also the consideration of specialized databases or alternative databases where data is echo'd to for the purposes of logging, metrics or reporting. Not to mention, certain layers of appropriate caching, which can still be less complex than some multi-database approaches.

[leafarlua]

What about the microservices/serverless functions world? This was another common topic over the years, that using SQL with this type of system was not optimal, I believe the issue being the connections to the SQL database and stuff.

[tracker1]

I think a lot of the deference to microservices/serverless is for similar reasons... you can work around some of this if you use a connection proxy, which is pretty common for PostgreSQL...

That said, I've leaned into avoiding breaking up a lot of microservices unless/until you need them... I'm also not opposed to combining CQRS style workflows if/when you do need micro services. Usually if you need them, you're either breaking off certain compute/logic workflows first where the async/queued nature lends itself to your needs. My limited experience with a heavy micro-service application combined with GraphQL was somewhat painful in that the infrastructure and orchestration weren't appropriately backed by dedicated teams leading to excess complexity and job duties for a project that would have scaled just fine in a more monolithic approach.

YMMV depending on your specific needs, of course. You can also have microservices call natural services that have better connection sharing heuristics depending again on your infrastructure and needs... I've got worker pools that mostly operate of a queue, perform heavy compute loads then interact with the same API service(s) as everything else.

[hunterpayne]

microservices are about the fact that administrative overhead for a software system increases exponentially w.r.t. the complexity of the system. Or to put it another way, microservices are a way to make a complex system without having the architecture explode in size. They have nothing to do with making more efficient software systems. They are about making complex systems that trade dev costs for operational costs.

[dalenw]

It's almost always a system design issue. Outside of a few specific use cases with big data, I struggle to imagine when I'd use NoSQL, especially in an application or data analytics scenario. At the end of the data, your data should be structured in a predictable manner, and it most likely relates to other data. So just use SQL.

[greenavocado]

System design issues are a product of culture, capabilities, and prototyping speed of the dev team

[mike_hearn]

Disclaimer: I work part time on the DB team.

You could also consider renting an Oracle DB. Yep! Consider some unintuitive facts:

• It can be cheaper to use Oracle than MongoDB. There are companies that have migrated away from Mongo to Oracle to save money. This idea violates some of HN's most sacred memes, but there you go. Cloud databases are things you always pay for, even if they're based on open source code.

• Oracle supports NoSQL features including the MongoDB protocol. You can use the Mongo GUI tools to view and edit your data. Starting with NoSQL is very easy as a consequence.

• But... it also has "JSON duality views". You start with a collection of JSON documents and the database not only works out your JSON schemas through data entropy analysis, but can also refactor your documents into relational tables behind the scenes whilst preserving the JSON/REST oriented view e.g. with optimistic locking using etags. Queries on JSON DVs become SQL queries that join tables behind the scenes so you get the benefits of both NoSQL and SQL worlds (i.e. updating a sub-object in one place updates it in all places cheaply).

• If your startup has viral growth you won't have db scaling issues because Oracle DBs scale horizontally, and have a bunch of other neat performance tricks like automatically adding indexes you forgot you needed, you can materialize views, there are high performance transactional message queues etc.

So you get a nice smooth scale-up and transition from ad hoc "stuff some json into the db and hope for the best" to well typed data with schemas and properly normalized forms that benefit from all the features of SQL.

[alexisread]

Good points, but Postgres has all those, along with much better local testing story, easier and more reliable CDC, better UDFs (in Python, Go etc.), a huge ecosystem of extensions for eg. GIS data, no licencing issues ever, API compatability with DuckDB, Doris and other DBs, and (this is the big one) is not Oracle.

[sgarland]

Unless I’ve missed something, Postgres doesn’t have automatic index creation, nor does it have JSON introspection to automatically convert it to a normalized schema (which is insane; I love it). It also doesn’t do any kind of sharding on its own, though of course forks like Citus exist. It definitely doesn’t do RAC / Exadata (not sure which part this falls under), where multiple nodes are connected and use RDMA to treat a bunch of SSDs as local storage.

I love Postgres, and am not a huge fan of Oracle as a corporation, but I can’t deny that their RDBMS has some truly astounding capabilities.

[alexisread]

I think that’s the beauty of PG here, you can find solutions to most of this:

Index creation https://stackoverflow.com/questions/23876479/will-postgresql...

JSON->DB schema https://jsonschema2db.readthedocs.io/en/latest/index.html

Pg shared disk failover is similar but RAC is quite unique, you’re not going to use though with a rented cluster?

https://www.postgresql.org/docs/current/different-replicatio...

Personally for me any technical advantages don’t outweigh the business side, YMMV :)

[mike_hearn]

RAC is a default part of any cloud Oracle DB, I think! I must admit I'm not an expert in all the different SKUs so there might be some that aren't, but if you rent an autonomous DB in the cloud you're probably running on ExaData/RAC. That's why the uptime is advertised as 99.95% even without a separate region.

> Index creation https://stackoverflow.com/questions/23876479/will-postgresql...

I was ambiguous. That's an answer telling how to create indexes manually, and saying that you get an index for primary keys and unique constraints automatically. Sure, all databases do that. Oracle can create arbitrary indexes for any relation in the background without it being requested, if it notices that common queries would benefit from them.

Forgetting to create indexes is one of the most common issues people face when writing database apps because the performance will be fine on your laptop, or when the feature is new, and then it slows down when you scale up. Or worse you deploy to prod and the site tanks because a new query that "works fine on my machine" is dog slow when there's real world amounts of data involved. Oracle will just fix it for you, Postgres will require a manual diagnosis and intervention. So this isn't the same capability.

> JSON->DB schema https://jsonschema2db.readthedocs.io/en/latest/index.html

Again I didn't provide enough detail, sorry.

What that utility is doing is quite different. For one, it assumes you start with a schema already. Oracle can infer a schema from a collection of documents even if you don't have one by figuring out which fields are often repeated, which values are unique, etc.

For another, what you get after running that utility is relational tables that you have to then access relationally via normal SQL. What JSON duality views give you is something that still has the original document layout and access mode - you GET/PUT whole documents - and behind the scenes that's mapped to a schema and then through to the underlying SQL that would be required to update the tables that the DB generated for you. So you get the performance of normalized relations but you don't have to change your code.

The nice thing about this is it lets developers focus on application features and semantics in the early stages of a startup by just reshaping their JSON documents at will, whilst someone else focuses on improving performance and data rigor fully asynchronously. The app doesn't know how the data is stored, it just sees documents, and the database allows a smooth transition from one data model to another.

I don't think Postgres has anything like this. If it does it'll be in the form of an obscure extension that cloud vendors won't let you use, because they don't want to/can't support every possible Postgres extension out there.

[rezonant]

You had me at "is not Oracle"

[tracker1]

I generally limit Oracle to where you are in a position to have a dedicated team to the design, deployment and management of just database operations. I'm not really a fan of Oracle in general, but if you're in a position to spend upwards of $1m/yr or more for dedicated db staff, then it's probably worth considering.

Even then, PostgreSQL and even MS-SQL are often decent alternatives for most use cases.

[mike_hearn]

That was true years ago but these days there's the autonomous database offering, where DB operations are almost all automated. You can rent them in the cloud and you just get the connection strings/wallet and go. Examples of stuff it automates: backups, scaling up/down, (as mentioned) adding indexes automatically, query plan A/B testing to catch bad replans, you can pin plans if you need to, rolling upgrades without downtime, automated application of security patches (if you want that), etc.

So yeah running a relational DB used to be quite high effort but it got a lot better over time.

[tracker1]

At that point, you can say the same for PostgreSQL, which is more broadly supported across all major and minor cloud platforms with similar features and I'm assuming a lower cost and barrier of entry. This is without signing with Oracle, Inc... which tends to bring a lot of lock-in behaviors that come with those feature sets.

TBF, I haven't had to use Oracle in about a decade at this point... so I'm not sure how well it competes... My experiences with the corporate entity itself leave a lot to be desired, let alone just getting setup/started with local connectivity has always been what I considered extremely painful vs common alternatives. MS-SQL was always really nice to get setup, but more recently has had a lot of difficulties, in particular with docker/dev instances and more under arm (mac) than alternatives.

I'm a pretty big fan of PG, which is, again, very widely available and supported.

[mike_hearn]

Autonomous DB can run on-premises or in any cloud, not just Oracle's cloud. So it's not quite the same.

I think PG doesn't have most of the features I named, I'm pretty sure it doesn't have integrated queues for example (SELECT FOR UPDATE SKIP LOCKED isn't an MQ system), but also, bear in mind the "postgres" cloud vendors sell is often not actually Postgres. They've forked it and are exploiting the weak trademark protection, so people can end up more locked in than they think. In the past one cloud even shipped a transaction isolation bug in something they were calling managed Postgres, that didn't exist upstream! So then you're stuck with both a single DB and a single cloud.

Local dev is the same as other DBs:

    docker run -d --name <oracle-db> container-registry.oracle.com/database/free:latest

See https://container-registry.oracle.com

Works on Intel and ARM. I develop on an ARM Mac without issue. It starts up in a few seconds.

Cost isn't necessarily much lower. At one point I specced out a DB equivalent to what a managed Postgres would cost for OpenAI's reported workload:

> I knocked up an estimate using Azure's pricing calculator and the numbers they provide, assuming 5TB of data (under-estimate) and HA option. Even with a 1 year reservation @40% discount they'd be paying (list price) around $350k/month. For that amount you can rent a dedicated Oracle/ExaData cluster with 192 cores! That's got all kinds of fancy hardware optimizations like a dedicated intra-cluster replication network, RDMA between nodes, predicate pushdown etc. It's going to perform better, and have way more features that would relieve their operational headache.

[danny_codes]

But then you’d have to interact with Oracle.

So.

Yeah no sane person would be that stupid

[OtomotO]

If you have an option, never ever use Oracle!

Never!

[freedomben]

I wanted to hate you for suggesting Oracle, but you defend it well! I had no idea

[ignoramous]

> One would think that for a startup of sorts, where things changes fast and are unpredictable, NoSQL is the correct answer. And when things are stable and the shape of entities are known, going for SQL becomes a natural path.

NoSQL is the "correct" answer if your queries are KV oriented, while predictable performance and high availability are priority (true for most "control planes"). Don't think any well-designed system will usually need to "graduate" from NoSQL to SQL.

Prior: https://news.ycombinator.com/item?id=22249490

[AlotOfReading]

There's plenty of middle ground between an unchanging SQL schema and the implicit schemas of "schemaless" databases. You can have completely fluid schemas with the full power of relational algebra (e.g. untyped datalog). You shouldn't be using NoSQL just because you want to easily change schemas.

[hunterpayne]

"NoSQL is the correct answer."

No, no it isn't. It never is. Just as building your house on a rubber foundation isn't the correct answer either. This is just cope. Unless your use cases don't care about losing data or data corruption at all, NoSQL isn't the correct answer.

[leafarlua]

You are probably correct. I was just parroting what I have seem from the industry. Isn't it common for even big start ups to have decided to go with rubber for their foundation? It seems to not be a wise decision, yet many engineers do take this path. And that is why I end up confused on this type of discussion.

[akdev1l]

> end up re-implementing relational databases at the application level anyway

This is by design, the idea is that scaling your application layer is easy but scaling your storage/db layer is not

Hence make the storage dumb and have the application do the joins and now your app scales right up

(But tbh I agree a lot of applications don’t reach the scale required to benefit from this)

[tshaddox]

I've never used DynamoDB in production, but it always struck me as the type of thing where you'd want to start with a typical relational database, and only transition the critical read/write paths when you get to massive scale and have a very good understanding of your data access patterns.

[icedchai]

Same. DynamoDB is almost never a good default choice unless you've thought very carefully about your current and future use cases. That's not to say it's always bad! At previous startups we did some amazing things with Dynamo.