Progress has "horizontal table partitioning" What do you think? --
But sharding goes a step furher
and could be good to have for us. At least oracle has it too:
Houtzager ICT consultancy & development
Progress has "horizontal table partitioning"
What do you think?
how would you use sharding if you had it?
btw: you can shard without having direct support for it in the database.
I would not use it until it became absolutely necessary, that is when extreme scalability is desired. There are con's. (Have I passed the test? ;-). My point is that it could be good for psc (and myself of course) to offer this feature, to be able to compete in a wider market. Concerning how I would use it: yes, I would have to learn.
i isn't a test. i am wanting to understand the nature of the problem you have to which sharding is the solution.
since your point is that it would mkae progress more competitive, it does not matter what the problem is. or put another way, sharding is for marketing.
It is not my intention to make a wordplay of this. I do not have a problem, but if the 'extreme scalability' is desired at sometime by one of my customers sharding could be a solution to that problem. So scalability can become a real and mattering problem. And not offering sharding could become one of the reasons a customers leaves the db. But never mind, no offense, keep up the good work and so forth. If I have critics they are much more addressed to psc marketing.
"Progress gives less priority to their engineering than they do to their marketing and acquisition.". Sorry, but I'm not a yesman.
well, i wasn't making a wordplay either. i was trying to gain some understanding of why you might need (as opposed to want) sharding. btw i did not mean to imply sharding was not useful or important. i agree we need to do better at marketing.
> i agree we need to do better at marketing.
You know, I have a share. I'm not young and still see openedge disappearing in Holland. And it's the trick I earn my money with.
One of the things one has to be aware of here is that company A claiming to have feature X is not equivalent to company B claiming to have feature X. This was clearly the case with DB encryption, for example where what PSC implemented was really quite different from what most other vendors offered under the same label. While the fact that they are different tends to mean that it is possible that one can make an argument for either implementation given the right circumstances, my sense on encryption was that PSC's implementation was more practically useful in the real world.
The same is clearly true of multi-tenant and I suspect the same is true of horizontal partitioning. So, before one gets excited about "PSC is terrible for not offering X", one should look at the choices with a more well rounded view.