Bug #14231 New Features and Concepts
Submitted: 22 Oct 2005 13:06 Modified: 21 Jul 2006 3:25
Reporter: Fabricio Mota Email Updates:
Status: Closed Impact on me:
Category:MySQL Server Severity:S4 (Feature request)
Version:Next OS:Any (Any)
Assigned to: CPU Architecture:Any

[22 Oct 2005 13:06] Fabricio Mota
Hi all,

I'm a system developer and projector, and faithful MySQL user, and I work with distributed systems there are some years. 

I have a project proposing the development of new features that could increase strongly efficience of the Enterprises' applications that uses MySQL Server, and also reduce their hardware costs many times. I think this can revolution the the typical database architectures, and put MySQL Servers many light-years beyond other database servers. 

It's about a new Engine able to perform transparent logistic distribution of data between many servers (that's not replication!). I think the development of this new engine could not be so expensive, comparated to the other mysql functionalities. The project propose is written in a paper, almost finished, described in a DOC file with approximately 25 pages.

Could any of you to be interested to, at least, take a look at it?
Can you give me an email address?

Thank you very much

How to repeat:
my e-mail address is fabricio.mota@gmail.com

Suggested fix:
Please don't leave this idea die!!!! It can really be very good.

Thank you very much!
[22 Oct 2005 14:56] MySQL Verification Team
Your idea is truly great, but can you tell us in few sentences in which way is your idea different from MySQL Cluster ??
[23 Oct 2005 2:20] Fabricio Mota
Sorry, at first I'm sorry to tell it for you, but I'm living a bad moment. I'm sad because a serious accident happened to my wife this week, and she is interned in an hospital. I'm praying and requesting God all the time to save her!!!

Now, about the proposed here:

Yes, I can. Here I propose a new Engine, that initially I call Distributed Data Engine - DDE. Comparating it with MySQL cluster, I get these differences:

Main objective
  Cluster: increase availability;
  DDE: increase applications performance;

  Cluster: Reducing costs of hardware and software infraestruture to warrant the service online (am I wrong?).
  DDE: Reducing costs of hardware infraestruture to increase machine performance and network traffic;

  Cluster: redundance (am I right?), warranting multiple points of service for fault tolerance;
  DDE: Logistic distribution of data, reducing network trafic and avoiding data centralization. Consider that each "server" used in a "DDE cluster" also could be a "MySQL cluster", for purposes of faults tolerance. The main objective of DDE is not to faults tolerance. Is to warrant to save data where it is better accessed.

A soon description:
Consider we have many distributed database servers, containing the same tables structures, but different data, like a national company with many offices, distributed along many provinces. Each province has its own data (different from others) saved on its own server. 
When a company wants to unify its data, it decides to centralize data in a server (or server cluster) that is centralized. It may causes an overloaded server and increase network around the server (or cluster). 
The idea is, at the same time, to distribute data logistically (the data will be where it is most accessed), but the cluster still warrant to work integrated for all network management and integrated operations, with transparence! 
Can I send this document for you to read at least the abstract and the introduction? I think it could be most comprehensive!

Thank you very much

Fabricio Mota
[23 Oct 2005 2:23] Fabricio Mota
Ops, where I said "increase performance and network traffic", I mean "increase performance and reduce network traffic"!
[25 Oct 2005 14:35] MySQL Verification Team
I am so sorry to hear that about your wife.

Please let us know how things evolve.

Regarding your concept, seems to me it is quite similar to our federated storage engine . 

What do you think ??

Also, if it is different, how would you implement your idea ??
[26 Oct 2005 1:30] Fabricio Mota

I thank you very much for care and solidarity about my wife. She is still in the hospital, but healing day by day, step by step. So we know we also may find true friends here. Thank you again.

Sinisa is your name? Where are you from?
I am from Brazil.

About FEDERATED Engine: it's a good idea, but FEDERATED and DDE (proposed here) have different purposes. Federated mean to delegate data from a table into a remote server. DDE mean to distribute data from a set of tables using logistic criteria.

I really don't know how to implement the idea yet, because I have no development team, neither I have MySQL project in hands (UML or something like this). Furthermore, to understant its (long) code has been "a little" hard... (lot of laughs).

So, I'm almost finishing a document containing this project, that is not done yet. But this document have 4 figures which explain very well the deep essence of this project. Could I send its previa to you, only for you spend your time seeing the 4 first figures (and read the abstract)? 

I promise you thats all!!

Thank you again. My regards,

Fabricio Mota
[28 Oct 2005 1:50] Fabricio Mota
Hi all,

Anybody there?
[21 Jul 2006 3:25] Jim Winstead
Closing the feature request, this is an external project now.

[14 Sep 10:45] Stefan Heisl
Hi Fabricio,

Your idea sounds intriguing, and it's evident that you're passionate about advancing MySQL Server capabilities. As any individual deeply invested in distributed systems, I can appreciate the practicable benefits of a new engine for logistic data distribution. Your imaginative and prescient for increasing efficiency and lowering hardware costs aligns with the ever-evolving demands of employer applications.

I'd be interested in exploring your proposal further. It's critical to consider how this new engine would integrate with present MySQL features and whether it may want to maintain the database's integrity while distributing facts across multiple servers. To foster interest from the community, you might consider sharing a summary or a teaser of your paper to pique curiosity.

Regarding contact, you can reach out to me here: https://container-news.com/supply-chain-reshaped-the-power-of-digital-transformation-in-lo... , I'm keen to delve into the details and share insights from my experience with distributed systems.