[mysql-dde] Re: implementation plan

  • From: "Fabricio Mota" <fabricio.mota@xxxxxxxxx>
  • To: <mysql-dde@xxxxxxxxxxxxx>
  • Date: Tue, 14 Mar 2006 19:24:39 -0300

Ah, ok. I understand it now. The idea is to produce experimental results to 
prove the model's viability. Right?

If so, I think it's possible.  :)

Furthermore, it could be good in different ways:

1 - Produce a prove to the model. Hey, I'll have to work! I'm gonna produce 
scientific theory about it (a Data Logistics Dissertation) and put the 
mathematical prove into that, and then send back to you validate / retify / 
increment. So, we will have the 3 pilars:
    a) The concept;
    b) experimental prove.
    c) The specification;

Good?

2 - Help the children to develop their ability, wich a new approach. They 
will get endorsement, and also take endorsement to us!

Se my answers below, too.

----- Original Message ----- 
From: "Peter B. Volk" <PeterB.Volk@xxxxxxx>
To: <mysql-dde@xxxxxxxxxxxxx>
Sent: Tuesday, March 14, 2006 6:16 AM
Subject: [mysql-dde] Re: implementation plan


> Hey,
>
> I would like to make clear that the implementation of the students is only 
> a
> prototype to prove that the theory works. Things like table masking etc. 
> are
> then guarantied to work properly. Also this implementation can be used for
> the production implementation as a skeleton.
>
> see the awnsers below.
>
> Peter
>
>> Maybe you're being ambitious, or otimist. I find this time very short to
>> guarantee these results.
>> How many people do you have to work on it?
>
> Two groups with 5 people. Both groups will produce a seperate solution.

Hey, we have a good team!!!! Very good!

>> Do all of you already know MySql architecture deep enough?
>
> no but the storage engines can be develped completly sperate to the MySql
> main application (see
> http://dev.mysql.com/doc/refman/5.1/en/custom-engine.html). This way they
> don't need to know the complete archetekture only a very smal part.
>> Don't you need to spend time studying and establishing project solution?
>
> Not realy. By giving them the smal steps for the implementation we take a
> way the workload for them of making a project plan.
>

I have a link to the MySql class diagram. I must be very useful tro help us 
to trace our project strategies. I will foward this email to you after this.

>>
>> Hey,
>>
>> How do you think people from MySQL AB (and from TU Dresden, too) are
> seeing
>> the "time" question? Are you being pressed by them in respect to it?
>
> They don't seem to feel under pressure. Also I'm not beeing presed into 
> this
> time frame. I just think that right here we have the possibility to have
> someone implement basic functionality to show to MySql that the concept it
> self actually works. Once the Students have completed their work we have a
> system where we can do insert, selects and deletes.

Ok. But please remember to:
1) See the class diagram to help us to take the shortest way; to avoid we 
"reinvent the wheel" :) ; and to get a implementation model the closest 
possible to MySql pattern model.
2) To ensure the development team to explore the "most definitive possible 
solution". So, we avoid re-works in forward.
3) Keep sending comunication about internal project to me :). That's because 
I have a "short" knowledge about design patterns in OO, so I could be 
helping you sometimes.
Maybe we could use this list (or a new one, if you prefer) to keep project 
comunication. What do you think?

Regards

FM

>> Consider the answers below:
>>
>> > Hey,
>> >
>> > had an apointment today with the TU. within the next 4 months we will
> have
>> > a
>> > prototype with the following functions:
>> >
>> > Add/remove server from cluster,
>>
>> I think this need a infraestructure implemented to save group management
>> data, able to guarantee the low-level comunication management. It spends
>> time...
>
> Yes this includes internal tables as proposed for the information_schema 
> in
> the design document.
>
>>
>> > Create/delete Cluster,
>>
>> Maybe it is a little bit faster, once the groups table being already
>> created.
>>
>> > Inserts/deletes and selects running,
>>
>> > Communication protocoll between Servers (the thing with the hint 
>> > message
>> > etc.),
>>
>> Traditionally, it could spend times, too... :)
>
> Well since we whant to map the communication to the traditional SQL we can
> have the Federated storage engine hacked. This engine already does the 
> exact
> same thing as we want to have. It take the primitive from above the 
> storage
> engine and then transforms this into an SQL for the remote table. So in
> theory this part is only a copy and paste job
>
>>
>> > "static" distribution of data amon servers (we limited the 
>> > deistribution
>> > to
>> > a hashcode of a cloumn that determans the location for the row)
>>
>> Well, if you consider punctual tests, in laboratory, it's ok. But if you
>> consider a complete test (from external user transaction), maybe this
> could
>> be a bit more complicated. This action maybe needs to have the engine
>> already working. What do you think?
>
> Well actually Mark (the guy from the TU doing this with me) whanted them 
> to
> implement real distribution already distributing the row by workload. By I
> thought that that would be overkill. Why would the engine need to be
> working? The problems with mutli user stuff is consistency guaranty. 
> Meaning
> Master election for p/f/u key stuff, Deadlock stuff or simular. So when 
> the
> basic single user system is there then the system would need to be 
> addabted
> to multiuser stuff.
>
>>
>> > Also as gloabl limitation: single user system.
>> >
>> > sounds ok to you?
>> >
>>
>> Sounds surreal. :)
>>
>> > Peter
>> >
>> >
>> > ----- Original Message ----- 
>> > From: "Fabricio Mota" <fabricio.mota@xxxxxxxxx>
>> > To: <mysql-dde@xxxxxxxxxxxxx>
>> > Sent: Monday, March 13, 2006 2:12 AM
>> > Subject: [mysql-dde] Re: implementation plan
>> >
>> >
>> >> I can't open the ftp...
>> >> An error occurs... :(
>> >>
>> >> ----- Original Message ----- 
>> >> From: "Peter B. Volk" <PeterB.Volk@xxxxxxx>
>> >> To: <mysql-dde@xxxxxxxxxxxxx>
>> >> Sent: Sunday, March 12, 2006 6:56 PM
>> >> Subject: [mysql-dde] implementation plan
>> >>
>> >>
>> >> > Hey,
>> >> > I've added the todo list in the global document repository and also
>> > added
>> >> > an intial version of the implementation plan. Within the
> implemntation
>> >> > plan the blue marked topics are candidates for the sudents to
>> >> > implement.
>> >> >
>> >> > Peter
>> >> > MySql-DDE discussion list
>> >> > www.freelists.org/
>> >> >
>> >>
>> >> MySql-DDE discussion list
>> >> www.freelists.org/
>> >>
>> >
>> > MySql-DDE discussion list
>> > www.freelists.org/
>> >
>>
>> MySql-DDE discussion list
>> www.freelists.org/
>>
>
> MySql-DDE discussion list
> www.freelists.org/
> 

MySql-DDE discussion list
www.freelists.org/

Other related posts: