Home
>
WinDev Forum
>
Topic
Performance of Hyper File Classic (compared to C/S)
Posted by Frank Missel
|
Frank Missel
Performance of Hyper File Classic (compared to C/S) March 28, 2008 06:58AM |
Hi,
I have a project which will mostly be used by a single or few users as well as run batch processing (ETL). Thus I have so far considered only using Hyper File Classic for that part of the application. However, since other projects of the same application will later use Hyper Files C/S I am considering to use that for all tables.
My question is if there is any performance gain of using Classic compared to the C/S since there is a lot database processing (ETL activities) in the above mentioned project. If there is not that much difference, I would be prone to just make the switch right away and use C/S for all tables in the database. However, sometimes C/S databases are considerable slower than direct file access databases.
Does anyone have experience with how performance compares?
Do the Hyper File C/S take advantage of multiple CPU's or cores on the same system so that you get parallel processing of queries or do you need to go with dedicated database systems like MS SQL Server, Oracle, etc. for that?
Can Hyper File C/S handle databases of say 3-500 Gb?
I know that in such databases performance depends very much on proper design of tables, keys, relations and the queries used. So my question is more the to general stability and performance of Hyper File C/S for databases of that size or perhaps even bigger.
What is the size of the largest HF C/S bases in operation?
Best regards,
Frank
I have a project which will mostly be used by a single or few users as well as run batch processing (ETL). Thus I have so far considered only using Hyper File Classic for that part of the application. However, since other projects of the same application will later use Hyper Files C/S I am considering to use that for all tables.
My question is if there is any performance gain of using Classic compared to the C/S since there is a lot database processing (ETL activities) in the above mentioned project. If there is not that much difference, I would be prone to just make the switch right away and use C/S for all tables in the database. However, sometimes C/S databases are considerable slower than direct file access databases.
Does anyone have experience with how performance compares?
Do the Hyper File C/S take advantage of multiple CPU's or cores on the same system so that you get parallel processing of queries or do you need to go with dedicated database systems like MS SQL Server, Oracle, etc. for that?
Can Hyper File C/S handle databases of say 3-500 Gb?
I know that in such databases performance depends very much on proper design of tables, keys, relations and the queries used. So my question is more the to general stability and performance of Hyper File C/S for databases of that size or perhaps even bigger.
What is the size of the largest HF C/S bases in operation?
Best regards,
Frank
|
Fabrice Harari
Re: Performance of Hyper File Classic (compared to C/S) March 28, 2008 09:08AM |
Hi Frank...
Classic and CS version of HF are based on the exact file structure (you can copy the files from one to the other). the only difference is that in classic you read/write in the file through your local HF dll, while in CS you have another layer doing so remotely using the HF dll over there
For a SINGLE user access there is therefore a small possibility that HF classic would be faster, as you remove the intermediate layer. Of course, in that case, the files would be local, so you would also remove the network transfer.
As soon as you have MORE than one user, HFCS is faster. The reason? As only ONE app access the files locally (I'm talking about the Server app, Manta), the files are blocked in exclusive mode.
On the contrary, if you would to access HF classic files through the network with several concurrent users, then each local HF dll would have to manage shared access, locking and releasing, etc...
Finally, as switching from one to the other is really a question of a few lines of codes, you can offer both possibilities in your app quite easily, so you can test on your own
Best regards
Fabrice Harari
WinDev, WebDev, WinDev Mobile Video Courses
Classic and CS version of HF are based on the exact file structure (you can copy the files from one to the other). the only difference is that in classic you read/write in the file through your local HF dll, while in CS you have another layer doing so remotely using the HF dll over there
For a SINGLE user access there is therefore a small possibility that HF classic would be faster, as you remove the intermediate layer. Of course, in that case, the files would be local, so you would also remove the network transfer.
As soon as you have MORE than one user, HFCS is faster. The reason? As only ONE app access the files locally (I'm talking about the Server app, Manta), the files are blocked in exclusive mode.
On the contrary, if you would to access HF classic files through the network with several concurrent users, then each local HF dll would have to manage shared access, locking and releasing, etc...
Finally, as switching from one to the other is really a question of a few lines of codes, you can offer both possibilities in your app quite easily, so you can test on your own
Best regards
Fabrice Harari
WinDev, WebDev, WinDev Mobile Video Courses
|
Frank Missel
Re: Performance of Hyper File Classic (compared to C/S) March 28, 2008 11:49AM |