Reaction against RDBMS?

[Aaron W. Hsu]

Lately there has been a lot of talk against RDBM systems. The NoSQL type
stuff seems to be getting really popular in some circles or has been for a
few years. I feel like this is an over-reaction and perhaps misinformed.
Being a casual user of RDBM systems in my own programming, and understanding
a few things about Database management, I fail to see the attraction of
these other Database replacements.

Is there a real, fundamental justification for this movement, or is it just
another fad that comes and goes around computer science in general?

--
Aaron W. Hsu | arc...@sacrideo.us | http://www.sacrideo.us
לֵ֤ב חֲכָמִים֙ בְּבֵ֣ית אֵ֔בֶל וְלֵ֥ב כְּסִילִ֖ים בְּבֵ֥ית שִׂמְחָֽה׃

[Robert Klemme]

On 01/21/2013 09:11 PM, Aaron W. Hsu wrote:
> Lately there has been a lot of talk against RDBM systems. The NoSQL type
> stuff seems to be getting really popular in some circles or has been for a
> few years. I feel like this is an over-reaction and perhaps misinformed.
> Being a casual user of RDBM systems in my own programming, and understanding
> a few things about Database management, I fail to see the attraction of
> these other Database replacements.
>
> Is there a real, fundamental justification for this movement, or is it just
> another fad that comes and goes around computer science in general?

Yes and yes.

robert

[Aaron W. Hsu]

Robert Klemme wrote:

> Yes and yes.

Would you mind articulating those reasons? I'd like to get a broader
perspective on the issue. Most of what I find is usually from someone saying
the equivalent of, "RDBMS stinks, of course X is better."

[Robert Klemme]

On 01/22/2013 05:54 AM, Aaron W. Hsu wrote:
> Robert Klemme wrote:
>
>> Yes and yes.
>
> Would you mind articulating those reasons? I'd like to get a broader
> perspective on the issue. Most of what I find is usually from someone saying
> the equivalent of, "RDBMS stinks, of course X is better."

Yes: not all applications need the strict consistency rules (ACID) which
typical RDBMS implementations of TX handling provide and not all
applications need immediate access to all basic data fields (i.e.
retrieval of complete documents is OK for many applications or even more
efficient). By abandoning these more performance can be achieved and
replication becomes easier just to name . So NoSQL database have real
benefits for particular use cases (same as RDBMS have as well - just
other use cases).

Yes: they are hyped which you can see any time when looking in the
media. When people start implementing RDBMS like TX handling on top of
NoSQL databases "by hand" you should be suspicious.

Cheers

robert

[Aaron W. Hsu]

Robert Klemme wrote:

> Yes: not all applications need the strict consistency rules (ACID) which
> typical RDBMS implementations of TX handling provide and not all
> applications need immediate access to all basic data fields (i.e.
> retrieval of complete documents is OK for many applications or even more
> efficient). By abandoning these more performance can be achieved and
> replication becomes easier just to name . So NoSQL database have real
> benefits for particular use cases (same as RDBMS have as well - just
> other use cases).

Okay, thanks for the analysis. How would you compare component files
(datatypes which are implicitly serialized to the filesystem but which
appear just as regular datatypes in the language once loaded), columnar
databases (such as VectorNova's RDBMS), and NoSQL sort of databases? I also
recall some databases going through a lot of effort to make XML handling go
hand-in-hand with the Relational model. I tend to think of XML objects as
complete, structured documents, so I immediately thought of that when you
mentioned the whole document retrieval method. Do you think this is a
misguided idea (mixing XML with the relational model)? Do you think it's a
completely different use case than what NoSQL is good for?

[Robert Klemme]

On 01/23/2013 03:05 AM, Aaron W. Hsu wrote:
> Robert Klemme wrote:
>
>> Yes: not all applications need the strict consistency rules (ACID) which
>> typical RDBMS implementations of TX handling provide and not all
>> applications need immediate access to all basic data fields (i.e.
>> retrieval of complete documents is OK for many applications or even more
>> efficient). By abandoning these more performance can be achieved and
>> replication becomes easier just to name . So NoSQL database have real
>> benefits for particular use cases (same as RDBMS have as well - just
>> other use cases).
>
> Okay, thanks for the analysis. How would you compare component files
> (datatypes which are implicitly serialized to the filesystem but which
> appear just as regular datatypes in the language once loaded), columnar
> databases (such as VectorNova's RDBMS), and NoSQL sort of databases?

I don't have enough experience with all of these to make an informed
comparison.

> I also
> recall some databases going through a lot of effort to make XML handling go
> hand-in-hand with the Relational model. I tend to think of XML objects as
> complete, structured documents, so I immediately thought of that when you
> mentioned the whole document retrieval method. Do you think this is a
> misguided idea (mixing XML with the relational model)? Do you think it's a
> completely different use case than what NoSQL is good for?

Well, as you said: RDBMS with XML support can really do both - tables
and structured documents. I haven't seriously worked with XML feature
in RDBMS to judge fitness of this approach. It does seem to be a
general approach to include non typical data types into RDBMS so there
seems to be some need for this in the market - or at least vendors
believe there is.

Kind regards

robert