Hacker News Reader

Comment by redwood | Hacker News Reader

This is an incredibly short-sighted piece for two reasons

1) the data challenges of yesterday's internet giants are the data problems for just about every 21st century enterprise tomorrow. We're at the start of / in the midst of an irreversible data explosion.

2) Established players not taking new technologies and upstarts seriously is never evidence of the threat not being real. We know that on the contrary our industry is characterized by constant disruption where established players for whatever reason do not consistently stay at the front line of innovation and tend to get leapfrogged. (as an analogy Google+'s lurch into social may be akin to Oracle's lurch into nosql. Does that mean social was not real?)

Bottom line: It all depends on what you're doing. But more and more of us will be dealing with more and more complex data in the years to come. Of that, I'm certain.

View on HN

I agree that's a bit short sighted, but I think it's one of perspective. The mistake most people make is to believe that Oracle (and other relational DB vendors) are competing directly against the various NoSQL stores (this includes doc stores, scalable key-value stores, etc.). I don't think that's true. I think that relational databases are really competing against industry-specific SaaS. So instead of implementing their own database for inventory & sales, companies may opt to use a service. These SaaS companies, in turn, are more likely to adopt a variety of technologies (including NoSQL and relational). Since the SaaS companies serve more than a single company, they're also more likely to adopt easily scalable technology like Cassandra. So fewer businesses will need to purchase databases (of any sort), but the ones that do will have greater scalability needs.

Agree. Where you see nonrelational popping up analogously is particularly in new SaaS shops, many of which aim to take some Oracle's pie.

NoSQL is absolutely competing with Oracle and other relational DB vendors. You are seeing it today with the valuations for companies like Mongo or DataStax. SaaS is hardly a big enough market compared to say every enterprise. And surely the companies listed on the client pages agree with me.

And people forget something important with SaaS. Data sovereignty. Here in Australia for example there are many enterprise companies who are forbidden from using ANYTHING that is hosted in the US due to the grey legal area e.g. Patriot Act. So in-house databases are absolutely still here to stay.

It's interesting, my experience with companies here in the U.S. (mainly SMBs) are that they feel much more comfortable using a SaaS compared to adopting an internal NoSQL store. Granted most of these companies aren't highly technical, but I suspect for them using a SaaS is easy and doesn't require much internal expertise. Adopting a new NoSQL store though may require training, etc. Of course I concede that things may be different in larger enterprises and companies outside the U.S.

I would bet, however, that most of the clients that do business with Mongo or DataStax still intend on using/maintaining a relational system. I haven't really encountered many companies that's decided to completely dump their relational systems in favor of something else.

NoSQL is not a superior tech, it's a trade-off.

What you are arguing is that future people will value data security and performance less than scalability... Well, I doubt it, but my crystal ball is as useless as yours.

I'm not saying it's superior, rather it finds a significant place in the stack for many of the most cutting edge data focused companies. So it's here to stay. My argument is that those cutting edge data focused companies will be just about everyone to varying degrees in the years to come. There's room for many technologies in the enterprise abd it's naive to assume as the author does that Facebook and Google's challenges are unicorns.

Facebook's and Google's challenges are not unicorns. But they are leviathans, and the author is somewhat correct to observe that NoSQL solves problems that we all "would love to have".

Or perhaps only think we have - I know a lot of developers who believe that an RDBMS should be kept at arm's length and only spoken to through an ORM, or who've never spent much time exploring what they're missing by never straying far from SQLite and MySql. These same folks seem to have a remarkable ability run into "big data" problems that "can't be solved by an ACID-compliant RDBMS" well short of the point where I'd only be thinking, "Huh, this DB's getting big enough that I might need to spend more time thinking about my indexing strategy."

I do agree wholeheartedly that more and more companies will be encountering situations where ACID proves to be a impediment to scalability in the future. But that's then, this is now, premature optimization is the root of all evil, and "Getting ready for a problem we might encounter in the coming years" is often just a fancy way of saying, "Solving a problem we don't actually have."

It's a superior tech because it allows you to develop to your business domain (which changes all the time anyway). Supporting infrastructure really should not be more important than your app code. SQL is almost a first class citizen in most old apps, which is tying dependencies too closely and makes for some bad apps.

Also for all consumer apps, people value scalability over security and performance today, yesterday, last week, a year, 5 years ago, whatever. You should be using NoSQL for that. And B2B almost all business domains.

Reply on news.ycombinator.com