[xquery-talk] Aggressivity and the Database Bubble WASRe: Linkedin humor for the weekend (2)

[daniela florescu]

Yes, Pavel, I know that. And you know that. And a bunch of real database people know that. There are plenty of us.

However, our voice gets lost ( and as Ihe said, we are more likely for us to be treated as crazy nuts then anything else….)
in face of the hundreds of millions of dollars that are being poured into marketing of those products.

One of my favorite is CouchDb marketing slogan: "N1QL is the FIRST TIME  in history when people can query the STRUCTURE
of the data.”

And yeah, MarkLogic’s CEO’s statement that MarkLogic invented a solution for solving heterogeneous data problem is another favorite of
mine too.

How does THAT look for you, after the past 25 years of semi-structured databases ?

I cannot do anything else then being stunned by how gullible and uneducated people are….

And wait for the bubble to pass…

BTW, an interesting related read:
https://www.linkedin.com/pulse/linkedin-has-become-wmd-philippe-collard?trk=hp-feed-article-title

Best
Dana



> On Jun 16, 2015, at 10:17 AM, Pavel Velikhov <pavel.velikhov at gmail.com> wrote:
> 
> Daniela,
> 
>   Its even worse than what you describe, because the vendors have to LIE about the properties of their systems, because they don’t have
> time to build reliable systems with all the guarantees that the marketing wants.
> 
>  Examples from the “database” world:
> 
>    - Cassandra for most of their life had DIRTY_WRITES as the highest isolation level. They lost data left and right, new releases would boast new features that were ALL unreliable. When confronted with tests, they accepted some into the test harness, but continued making claims that were so easy to break. And then they wrote a driver for Spark - which was completely broken. That said, at first they claimed to be a universal system with all sorts of workloads. Then, as they stopped scaling beyond simple INSERT/SELECT use-cases, they claimed that’s what the initial design was about. 
> 
>    - Elastic Search loses data left and right, had a bug in their consensus mechanism for a long time, claim to have included Jepsen tests in their harness, but didn’t really fix the bugs that Jepsen reported. Documentation for this project is a must read on how not to document your system.
> 
>    - Mongo was breaking left and right on pretty simple queries. Also have a data loss problem. And a bunch of other problems. But we already have our favourite punch lines about Mongo :)
> 
> Examples from the “Hadoop” world:
> 
>   - My favourite here is using Hadoop for simple massively parallel workloads, where it performed hundreds of times slower than scripts. With GNU_Parallel out there 90% of current Hadoop jobs can be done faster, will run faster and require much less monitoring than a hadoop cluster.
> 
> So hope the marketing bubble bursts soon and hope that high quality systems will reemerge.
> 
>> On 16 Jun 2015, at 19:26, daniela florescu <dflorescu at me.com> wrote:
>> 
>>>> 
>>> 
>>> Agreed, the database software industry is about the most vicious example of naked aggressive capitalism that I have seen anywhere.
>>> 
>>> But I don’t think it has got worse since Oracle started trying to wipe out their competitors in the mid 80s. It has always been that bad.
>> 
>> 
>> Michael,
>> 
>> believe it or not, I think it got worse since the days of Larry Ellison (as bas and aggressive as it was… :-)
>> 
>> The first generation of databases grew organically, with their customer base … they were busy fighting SPLITTING an exiting market
>> which was naturally growing. Those databases were DB2(IBM), Oracle, SQL Sever. None of them had a VC behind it….
>> 
>> The new generation of databases (Cloudera, DataStax, Mongo. CouchDB. MarkLogic…etc) are NOT growing organically.
>> 
>> They are all financed by Venture Capitalists. They all took between 100M and 200M, sometimes more,  investment money from VCs. 
>> (And I can tell you, lending money from VCs is worse then lending money from the mafia….. if you don’t give it back… they’ll find you ….)
>> 
>> A VC naturally wants his investment returned 50X (or whatever X they want) in a fixed amount of time (2-5 years, or whatever).  This is how VC world works.
>> 
>> So…. this new generation of databases, being financed by VCs,  CANNOT grow naturally and organically with the market…..
>> 
>> Their growth speed is imposed by the VCs, and not by the market growth.
>> 
>> They have to pull customers out of their a..s. They have to create artificial customers.
>> 
>> They have to go to each other’s throat for the meager number of customers.
>> 
>> Hence the general hysteria.
>> 
>> ===========
>> 
>> Hence all the horrible things that happens right now in the “database” industry, marketing screams all over the place, idiotic marketing messages 
>> (scale to the level of the “universe”..), bogus benchmarks, query languages that don’t NEED a specification, proprietary syntaxes to cover an existing standard
>> — because a standard would reduce the value of the company— bullying every single blogger in the industry to say what you want, physically
>> abusing people who dare to say something else, bribing of officials of all kinds…..
>> 
>> Gold rush, here we come again.
>> 
>> Science ( temporarely I hope ), left this field. 
>> 
>> How good a product is irrelevant right now. You can see that by watching the amount of money spent by this generation of databases in marketing and sales vs. engineering.
>> Usually it’s 10X. This was not true for Oracle, even if they did spend a large amount on sales.
>> 
>> My hope is that when the database bubble will crash, soon, VCs will finally get disappointed, and finally move away to another field, like locusts, so
>> we can come back and bring some scientific interest into this field.
>> 
>> But, yeah, I’ve never seen ANYTHING like what’s going on right now with the database companies in Sillicon Valley…..
>> 
>> Best
>> Dana
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> _______________________________________________
>> talk at x-query.com
>> http://x-query.com/mailman/listinfo/talk
> 
> С уважением,
> Павел Велихов
> pavel.velikhov at gmail.com
> 
> 
> _______________________________________________
> talk at x-query.com
> http://x-query.com/mailman/listinfo/talk