Wrox Home  
Search P2P Archive for: Go

  Return to Index  

asp_components thread: Why and When use COM???


Message #1 by walter@w... on Thu, 1 Nov 2001 14:59:36
Can someone explain when i shut use COM and when not?
Message #2 by Robert Chartier <rchartierh@a...> on Thu, 01 Nov 2001 09:18:21 -0800

all you need to do is keep in mind one thing:



ASP is a hack.



What i mean by that is that it has such little support for really 

everything.  Sure you can get a COM here and there to do your things, but 

really its pretty useless.  Dont worry, so is PHP.  JSP is much better, for 

every reason.  Then we get into .NET (DONE RIGHT), it is a huge upgrade 

(better solution) than most realize.



Consider this...



http://www.15seconds.com/issue/011023.htm



COM could exists at pretty much at all layers.



But realistically, consider it for these layers:



Distributed Logic

Proxy Tier

Business Tier

Data Access Tier





....discuss more...





/rob













At 02:59 PM 11/1/2001 +0000, you wrote:

>Can someone explain when i shut use COM and when not?

>








Robert Chartier

Author and Developer

xxx-xxx-xxxx

rob@a...

http://www.aspfree.com/devlinks

http://www.aspfree.com/authors/robert

http://www.aspalliance.com/nothingmn

http://www.15seconds.com/issue/wa.htm#chartier



Message #3 by "Lewis, James" <James.Lewis@i...> on Thu, 1 Nov 2001 10:40:11 -0700
Personally I incorporate my own COM controls for the following reasons:



- I need the performance from a compiled component as compared to scripted

hosts such as IIS

	o I have a control that generates several thousand .ini files based

on some database settings. This control uses the direct win32 API file

generation (written in C++/ATL). Just by moving away from the file-system

object we realized a 2x speed increase. Improvements in the algorithms

realized yet another 2x increase. This environment also allows me to utilize

threads.



- I need hyper close control over security, such as instantiation a DCOM

control under specific permissions on specific machines.

	o I wrote an ATL control that instantiates objects on remote

machines. The control takes a user ID, domain, and password which guarantees

that the SCM will not receive a NULL ACL on the remote machine. This has

been highly useful for me when I wanted to restrict a control that needs

full file system access to only one share.



- I need a highly distributed DCOM environment that is more sophisticated

that IIS or DCOMCNFG could hope to achieve.

	o There's only so much DCOMCNFG can do - for more sophisticated DCOM

environments, such as say a custom load balancing environment that

piggybacks on the Windows2000 browser for location of remote servers, a

scripted environment is hopeless.



- I need access to an O/S call or other low-level service that can only be

obtained at the Win32 level

	o I have a hyper file-copy control that uses multiple threads an

overlapped I/O to copy up to 10,000 files to a distribution server. The

control is triggered from IIS but internally uses low-level API's including

overlapped I/O, thread pooling, etc.





J



-----Original Message-----

From: walter@w... [mailto:walter@w...]

Sent: Thursday, November 01, 2001 8:00 AM

To: ASP components

Subject: [asp_components] Why and When use COM???





Can someone explain when i shut use COM and when not?

 




$subst('Email.Unsub')

Message #4 by "Michael Paul ( LTI )" <MichaelP@l...> on Thu, 1 Nov 2001 14:17:59 -0800
Kinda Off the Subject but...



I just heard from a friend that he was told that the .NET runtime was faster

then COM/ATL. He heard this at the .NET show-tell thing in LA last week. Has

anyone heard this as well??. I would find this ALMOST impossible. He also

said

that the Microsoft was moving away from COM+ in the .NET architecture. 







-----Original Message-----

From: Lewis, James [mailto:James.Lewis@i...]

Sent: Thursday, November 01, 2001 9:40 AM

To: ASP components

Subject: [asp_components] RE: Why and When use COM???





Personally I incorporate my own COM controls for the following reasons:



- I need the performance from a compiled component as compared to scripted

hosts such as IIS

	o I have a control that generates several thousand .ini files based

on some database settings. This control uses the direct win32 API file

generation (written in C++/ATL). Just by moving away from the file-system

object we realized a 2x speed increase. Improvements in the algorithms

realized yet another 2x increase. This environment also allows me to utilize

threads.



- I need hyper close control over security, such as instantiation a DCOM

control under specific permissions on specific machines.

	o I wrote an ATL control that instantiates objects on remote

machines. The control takes a user ID, domain, and password which guarantees

that the SCM will not receive a NULL ACL on the remote machine. This has

been highly useful for me when I wanted to restrict a control that needs

full file system access to only one share.



- I need a highly distributed DCOM environment that is more sophisticated

that IIS or DCOMCNFG could hope to achieve.

	o There's only so much DCOMCNFG can do - for more sophisticated DCOM

environments, such as say a custom load balancing environment that

piggybacks on the Windows2000 browser for location of remote servers, a

scripted environment is hopeless.



- I need access to an O/S call or other low-level service that can only be

obtained at the Win32 level

	o I have a hyper file-copy control that uses multiple threads an

overlapped I/O to copy up to 10,000 files to a distribution server. The

control is triggered from IIS but internally uses low-level API's including

overlapped I/O, thread pooling, etc.





J



-----Original Message-----

From: walter@w... [mailto:walter@w...]

Sent: Thursday, November 01, 2001 8:00 AM

To: ASP components

Subject: [asp_components] Why and When use COM???





Can someone explain when i shut use COM and when not?

 




$subst('Email.Unsub')



 




$subst('Email.Unsub')

Message #5 by "Lewis, James" <James.Lewis@i...> on Fri, 2 Nov 2001 09:49:51 -0700
Yeah. My take on this is that .NET is partially advertising - if you look up

lots of the .NET documentation it's just re-packaged COM and Win32

interfaces. I'm betting that what this means is .NET is providing easier

access to the underlying infrastructure, but the underlying technology,

calling what you like, is still RPC or COM or whatever. For example, the RPC

interface to services is still the same as it was back in NT4.... and deep

down under the covers where nobody wants to look, message pumps are alive

and well  :o)



-----Original Message-----

From: Michael Paul ( LTI ) [mailto:MichaelP@l...]

Sent: Thursday, November 01, 2001 3:18 PM

To: ASP components

Subject: [asp_components] RE: Why and When use COM???





Kinda Off the Subject but...



I just heard from a friend that he was told that the .NET runtime was faster

then COM/ATL. He heard this at the .NET show-tell thing in LA last week. Has

anyone heard this as well??. I would find this ALMOST impossible. He also

said

that the Microsoft was moving away from COM+ in the .NET architecture. 







-----Original Message-----

From: Lewis, James [mailto:James.Lewis@i...]

Sent: Thursday, November 01, 2001 9:40 AM

To: ASP components

Subject: [asp_components] RE: Why and When use COM???





Personally I incorporate my own COM controls for the following reasons:



- I need the performance from a compiled component as compared to scripted

hosts such as IIS

	o I have a control that generates several thousand .ini files based

on some database settings. This control uses the direct win32 API file

generation (written in C++/ATL). Just by moving away from the file-system

object we realized a 2x speed increase. Improvements in the algorithms

realized yet another 2x increase. This environment also allows me to utilize

threads.



- I need hyper close control over security, such as instantiation a DCOM

control under specific permissions on specific machines.

	o I wrote an ATL control that instantiates objects on remote

machines. The control takes a user ID, domain, and password which guarantees

that the SCM will not receive a NULL ACL on the remote machine. This has

been highly useful for me when I wanted to restrict a control that needs

full file system access to only one share.



- I need a highly distributed DCOM environment that is more sophisticated

that IIS or DCOMCNFG could hope to achieve.

	o There's only so much DCOMCNFG can do - for more sophisticated DCOM

environments, such as say a custom load balancing environment that

piggybacks on the Windows2000 browser for location of remote servers, a

scripted environment is hopeless.



- I need access to an O/S call or other low-level service that can only be

obtained at the Win32 level

	o I have a hyper file-copy control that uses multiple threads an

overlapped I/O to copy up to 10,000 files to a distribution server. The

control is triggered from IIS but internally uses low-level API's including

overlapped I/O, thread pooling, etc.





J



-----Original Message-----

From: walter@w... [mailto:walter@w...]

Sent: Thursday, November 01, 2001 8:00 AM

To: ASP components

Subject: [asp_components] Why and When use COM???





Can someone explain when i shut use COM and when not?

 




$subst('Email.Unsub')



 




$subst('Email.Unsub')



 




$subst('Email.Unsub')


  Return to Index