The NoSql squad is on the loose again, therefore it is time once again to slay that particular dragon. Andy Oram, of O'Reilly fame writes about a recent NoSql conference (prior to it). The usual suspects are listed.
The hallmark (or carbuncle) of the NoSql "databases" is that they have no structured access, no transactional support, and no data reduction. You get all those, and more, with relational (sql, mostly) databases. What these NoSql (Oram's list: Cassandra, CouchDB, HBase, HypergraphDB, Hypertable, Memcached, MongoDB, Neo4j, Riak, SimpleDB, Voldemort) datastores do is just what files did with your grandfather's COBOL code; store data in an application specific format. You can browse through them at your leisure.
What they have in common is the use of key/value pairs for data. The proponents fail to comprehend that their "data" is just what an index is in a relational database. And it's nothing new. Back in the 1960's random access was supported by the "fully inverted file" paradigm. Here's what Joe Celko has to say ("Joe Celko's Data and Databases"): "In an inverted file, every column in a table has an index on it. It is called an inverted file structure because columns become files." Nothing new here, please move on. Gad, these young-uns persist in re-inventing square wheels.
IFF AI, as currently implemented, can intuit Special Relativity from just a clock tower and a moving tram will I take AI seriously.
Infinite granularity yields infinite complexity.
Leading an Insurrection is now an Official Act
The world is not linear.
-- Dr. McElhone/1974
Well, I'm going to do it, whether the women like it or not. I'm going to protect them.
-- The Orange Shitgibbon©anon./2024 [y'all get burkhas]
Scientists aren't vocal enough about science. There are large groups of people who think scientists are all frauds and who don't
believe in science, and they're being cultured by some of our far right-wing politicians, religious leaders and community leaders. -- Dr. McElhone/1974
Well, I'm going to do it, whether the women like it or not. I'm going to protect them.
-- The Orange Shitgibbon©anon./2024 [y'all get burkhas]
-- Drew Weissman/2022
To date, Microsoft is stating that organizations testing In-Memory OLTP have
seen transaction speeds improve by up to 30 times compared to past
performance, with the best performance gains achieved when the business
logic resides in the database and not in the applications.
-- Jonathan Watts/2015 [my emphasis]
-- Jonathan Watts/2015 [my emphasis]
I have had to explain and re-explain and re-explain and re-explain, you know, how relational databases work, what is an eigenvector,
what is dimensionality reduction.
-- Christopher Wylie/2018
... but Flash-based storage has such a different performance profile from rotating media, that I suspect that it will end up having
a large impact on filesystem design. Right now, most filesystems tend to be designed with the latencies of rotating media in mind.
-- Linus Torvalds/2007
-- Chris Date/2009
Election day's thought
Democracy is the theory that the common people know what they want, and deserve to get it good and hard. -- H. L. Menken/1915 [
See you next week in a brand new show©Heckle and Jeckle
In a time of SSD, multi-core/processor, two terabyte memory and Optane App Direct Mode machines, there is no reason not to
build from BCNF data. Time to do what Dr. Codd demonstrated. Technology has finally caught up with the maths.
About
- Robert Young
- Been doing databases, both relational and others, as well as operations research since Hector was a Pup.
Shameless Plug
I know a little something about SQL, DB2 (preferably LUW, z/OS if the money's right), Oracle, SQL Server, Postgres, and database design in general. Mentoring/teaching/consulting on a short term basis, too. And a bit of stat analysis if it's R.
gnuoytr at rcn dot com.
gnuoytr at rcn dot com.
5 comments:
Found a somewhat interesting link on a (the?) Cassandra promotion site via "hacker news" today:
https://www.cloudkick.com/blog/2010/mar/02/4_months_with_cassandra/
(relevant to me since they switched from PostgreSQL)
Related to what you have been promoting, this seemed to be a particularly relevant quote:
>> ... we faced issues of needing a much bigger machine and faster rotational disks ...
i.e. - if your "disks" don't rotate, would that be fast enough?
It's interesting to note that they have some SQL systems still, but they are MySQL. This suggests to me that this is their default "database", but they were forced to use PostgreSQL since MySQL does not scale well on multiprocessor machines.
Back 10 years ago or so when I was "shopping" for an open source database, I compared MySQL to PostgreSQL. MySQL was really scary (how about "ROLLBACK is not implemented", for starters) at the time. Most of the "ACID" stuff was left out then (around 2001). The MySQL team gradually added ACID support, and other extensions that PostgreSQL already had, but they pretty much lost all credibility in my eyes from then on. I've already done my share of dBASE / XBase apps in the late 80s. Slapping SQL on top of pretty much the same kind of technology is just making things worse, in my mind.
Along comes the Cassandra project: Oracle buys ... the MySQL support team, and many of them leave. They start a child project to strip down the current version of MySQL even more.
Why do I want this???
OK, so it's fun to preach to the choir. Any thoughts on this (particular link and/or Cassandra)?
Some empirical fuel for the fire:
http://www.yafla.com/dforbes/The_Impact_of_SSDs_on_Database_Performance_and_the_Performance_Paradox_of_Data_Explodification/
This guy puts together a test of one of these "NoSQL" claims. The target site claimed big improvements when they switched from MySQL (why is it always MySQL???). So this guy rigs up a simulation using humble SQL-Server (not my cup of tea, but that's the point -- it's pretty run of the mill), and stomps on their stats.
Then, he sets up an SSD and blows off the doors. Check it out for the numbers and methodology.
I'll spend some time tonight with Cassandra and the test. Been a bit busy trading.
Well, Cassandra looks a lot like dBaseII. That's not a complement. Once again, moving data control away from the data and back to the siloed application. Very Carnaby Street.
What's so infuriating about such zealots is that they come, by and large, from the OO world. And the prime directive of the OO world is that method and data are co-located, an idea which Dr. Codd was the first to define.
As to Dennis Forbes, I've added his site to Good Stuff. While he says he is not a DB guy, he shows more insight than I've usually come across in a member of the Younger Generation. Recommended.
Post a Comment