Re: A question of federation - was: Planning kernel release 2.0

2007-03-26 Thread Simon Laws

On 3/22/07, Jim Marino <[EMAIL PROTECTED]> wrote:


>> > Hi Meeraj
>>
>> From my perspective having demonstrable code in June would be spot
>> on as
>> I have to speak on SCA then and would consider a demo if we could
>> do it.
>>
Maybe we can even get something earlier -  I'm also speaking at JavaOne.

>> I don't have the knowledge yet to comment on the details of your
>> proposal just yet (hence the new subject) but a question. From a
>> future
>> demo point of view I would like to show various runtime options
>> some of
>> which are not federated  examples some of which are. Can I miss
>> out the
>> federation bit if I want to? For example, I would potentially like to
>> show a variety of scenarios
>>
>> - Hello world. the simplest possible single process example to get
>> people into how SCA works
>> - Standalone domain (a single VM)
>>  service provision (perhaps an AJAX style example where an SCA
>>

+1

I think it would be good to finish out some of the programming model
for web apps, in particular allowing injection on Services. If you
want to demo something with Flex, let me know as I have some contacts
there that may be able to help us.

For a simple example, in my talk I used the loan-app example from the
core samples. I basically started with a simple LoanClient that
talked to a LoanService (pulled from the core-samples) in a stateless
manner. I showed how components can be simple POJOs with no
annotations and then demonstrated how they are configured in a very
simple assembly. I've done a number of SCA presentations and the
approach that seems to resonate the most is first providing a high-
level picture of what SCA is by comparing it to Microsoft WCF. I then
have one slide basically saying "components offer services and have
references". This is similar to Don Box's description of WCF when he
talks about services having "ABCs". Right after that slide, I jump
into a POJO example with the point of saying "you [the audience]
already know how to write components in Java using SCA". I usually
show the assembly SCDL next which is just a few lines. Then I update
the example to apply conversational scope and callbacks to
demonstrate why the development paradigm is an evolution from what
current exists. At that, point I then say "this isn't something
entirely new, this can be done today...what is new is federation,
runtime selection of bindings, etc. etc."


>> composite provides services to the browser)
>>  service consumption (backend service access providing content
>> to my
>> AJAX service)
>> - Federated domain (multiple VM)
>>  How SCA describes many connected composites.
>>
>> I'm just starting now to look at how all the kernel stuff works so I
>> expect all this will become clear soon enough (I found your previous
>> posts giving explantion b.t.w - so am starting from there)

Cool. I think making this stuff a reality and moving it beyond
demoware would be a great way to promote collaboration. Let me know
how I can help.

Jim



-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Thanks for that Jim, makes life a little easier when someone has thought

through this before. I like the build up approach. Something I would also
like to do is include some alternative component type implementations.
Anyhow, it's a little while off yet so I'll see what comes together over the
coming two months. I'm less concerned now that we won't have anything to
show!

Regards

Simon


Re: A question of federation - was: Planning kernel release 2.0

2007-03-22 Thread Jim Marino

> Hi Meeraj

From my perspective having demonstrable code in June would be spot  
on as
I have to speak on SCA then and would consider a demo if we could  
do it.



Maybe we can even get something earlier -  I'm also speaking at JavaOne.


I don't have the knowledge yet to comment on the details of your
proposal just yet (hence the new subject) but a question. From a  
future
demo point of view I would like to show various runtime options  
some of
which are not federated  examples some of which are. Can I miss  
out the

federation bit if I want to? For example, I would potentially like to
show a variety of scenarios

- Hello world. the simplest possible single process example to get
people into how SCA works
- Standalone domain (a single VM)
 service provision (perhaps an AJAX style example where an SCA



+1

I think it would be good to finish out some of the programming model  
for web apps, in particular allowing injection on Services. If you  
want to demo something with Flex, let me know as I have some contacts  
there that may be able to help us.


For a simple example, in my talk I used the loan-app example from the  
core samples. I basically started with a simple LoanClient that  
talked to a LoanService (pulled from the core-samples) in a stateless  
manner. I showed how components can be simple POJOs with no  
annotations and then demonstrated how they are configured in a very  
simple assembly. I've done a number of SCA presentations and the  
approach that seems to resonate the most is first providing a high- 
level picture of what SCA is by comparing it to Microsoft WCF. I then  
have one slide basically saying "components offer services and have  
references". This is similar to Don Box's description of WCF when he  
talks about services having "ABCs". Right after that slide, I jump  
into a POJO example with the point of saying "you [the audience]  
already know how to write components in Java using SCA". I usually  
show the assembly SCDL next which is just a few lines. Then I update  
the example to apply conversational scope and callbacks to  
demonstrate why the development paradigm is an evolution from what  
current exists. At that, point I then say "this isn't something  
entirely new, this can be done today...what is new is federation,  
runtime selection of bindings, etc. etc."




composite provides services to the browser)
 service consumption (backend service access providing content  
to my

AJAX service)
- Federated domain (multiple VM)
 How SCA describes many connected composites.

I'm just starting now to look at how all the kernel stuff works so I
expect all this will become clear soon enough (I found your previous
posts giving explantion b.t.w - so am starting from there)


Cool. I think making this stuff a reality and moving it beyond  
demoware would be a great way to promote collaboration. Let me know  
how I can help.


Jim



-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: A question of federation - was: Planning kernel release 2.0

2007-03-22 Thread Simon Laws

On 3/22/07, Jeremy Boynes <[EMAIL PROTECTED]> wrote:


On Mar 22, 2007, at 8:47 AM, Simon Laws wrote:
>> Ok, cool, so I can run a simple app in a single VM. Let me try it
>> out.

Just to set expectations, I don't think the system configuration in
the default runtime has been switched over to the federated deployer
yet. So if you run the calc sample on that runtime then it will still
be using the old stuff.

If you want to experiment with the federated stuff, you would need to
use the scdl from the master profile in the demo.

--
Jeremy


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

so I get the option to include/exclude federation by modifying the system

configuration? Nice. Not saying I don't want federation but nice to have the
option.

S


Re: A question of federation - was: Planning kernel release 2.0

2007-03-22 Thread Jeremy Boynes

On Mar 22, 2007, at 8:47 AM, Simon Laws wrote:
Ok, cool, so I can run a simple app in a single VM. Let me try it  
out.


Just to set expectations, I don't think the system configuration in  
the default runtime has been switched over to the federated deployer  
yet. So if you run the calc sample on that runtime then it will still  
be using the old stuff.


If you want to experiment with the federated stuff, you would need to  
use the scdl from the master profile in the demo.


--
Jeremy


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: A question of federation - was: Planning kernel release 2.0

2007-03-22 Thread Simon Laws

On 3/22/07, Meeraj Kunnumpurath <[EMAIL PROTECTED]> wrote:


Simon,

As Jim mentioned in an earler email, A single-VM or runtime physical
topology is just a degenerate case of a multi-VM model. In a single-VM
scenario there is only one profile that runs the master and the slave.
With some minor modifications, we should be able to run the TSS demo in
a single VM mode from deploying the SCDL within the admin console,
through to invoking the application. That has been one of the key
drivers for separating logical and physical aspects of the model.
Physical model is generated from the logical model based on the targeted
runtimes for the components included in the composite that is being
contributed. In a single-VM model, all the physical components will be
targeted onto the same runtime.

In fact, though the demo showed the federation aspects of Tuscany, the
actual application after deployment onto the slave ran in single-VM mode
using local bindings.

HTH

Ta
Meeraj

-Original Message-
From: Simon Laws [mailto:[EMAIL PROTECTED]
Sent: Thursday, March 22, 2007 12:22 PM
To: tuscany-dev@ws.apache.org
Subject: A question of federation - was: Planning kernel release 2.0

On 3/22/07, Meeraj Kunnumpurath <[EMAIL PROTECTED]> wrote:
>
> Hi,
>
> Now that the SPI is getting stable and we have the initial end-to-end
> story for federation working, I would suggest we plan for the final
> release for kernel 2.0, with emphasis on federation and user
experience.
> I was thinking about aiming for a beta in June in time for TSSJS
> Barcelona and the final release for August. Maybe we can have couple
> of alpha releases from now and June as well. These are the features, I

> would like to see in 2.0.
>
> 1. Tidy up anything required in physical model, now that it is
> starting to take good shape.
> 2. Tidy up generators from logical to physical model.
> 3. Fix the JXTA discovery issues, also investigate other discovery
> protocols.
> 4. Federation end-to-end fully completed, this would include, maybe,
> profiles advertising their capabilities and the information being used

> in intent-based autowiring etc.
> 5. Intent-based auto wiring
> 6. Emphasis on end user experience in terms of ease of use.
> 7. Assembly service, this kind of now related to the generators that
> have been introduced in the last week or so 8. Artifact management,
> especially mobile code when we target components to remote profiles.
>
> Also, now the SPI has started settling in, we need to start looking at

> binding and container extensions as well. Some of the bindings I would

> be interested in are,
>
> 1. JMS
> 2. AMQP
> 3. Hessian
>
> Ta
> Meeraj
>
>
> *
>
> You can find us at www.voca.com
>
> *
> This communication is confidential and intended for the exclusive use
> of the addressee only. You should not disclose its contents to any
> other person.
> If you are not the intended recipient please notify the sender named
> above immediately.
>
> Registered in England, No 1023742,
> Registered Office: Voca Limited
> Drake House, Three Rivers Court,
> Homestead Road, Rickmansworth,
> Hertfordshire, WD3 1FX. United Kingdom
>
> VAT No. 226 6112 87
>
>
> This message has been checked for all email viruses by MessageLabs.
>
> -
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
> Hi Meeraj

From my perspective having demonstrable code in June would be spot on as
I have to speak on SCA then and would consider a demo if we could do it.

I don't have the knowledge yet to comment on the details of your
proposal just yet (hence the new subject) but a question. From a future
demo point of view I would like to show various runtime options some of
which are not federated  examples some of which are. Can I miss out the
federation bit if I want to? For example, I would potentially like to
show a variety of scenarios

- Hello world. the simplest possible single process example to get
people into how SCA works
- Standalone domain (a single VM)
 service provision (perhaps an AJAX style example where an SCA
composite provides services to the browser)
 service consumption (backend service access providing content to my
AJAX service)
- Federated domain (multiple VM)
 How SCA describes many connected composites.

I'm just starting now to look at how all the kernel stuff works so I
expect all this will become clear soon enough (I found your previous
posts giving explantion b.t.w - so am starting from there)

Regards

Simon


This message has been checked for all email viruses by MessageLabs.

This message has been checked f

RE: A question of federation - was: Planning kernel release 2.0

2007-03-22 Thread Meeraj Kunnumpurath
Simon,

As Jim mentioned in an earler email, A single-VM or runtime physical
topology is just a degenerate case of a multi-VM model. In a single-VM
scenario there is only one profile that runs the master and the slave.
With some minor modifications, we should be able to run the TSS demo in
a single VM mode from deploying the SCDL within the admin console,
through to invoking the application. That has been one of the key
drivers for separating logical and physical aspects of the model.
Physical model is generated from the logical model based on the targeted
runtimes for the components included in the composite that is being
contributed. In a single-VM model, all the physical components will be
targeted onto the same runtime.

In fact, though the demo showed the federation aspects of Tuscany, the
actual application after deployment onto the slave ran in single-VM mode
using local bindings.

HTH

Ta
Meeraj

-Original Message-
From: Simon Laws [mailto:[EMAIL PROTECTED] 
Sent: Thursday, March 22, 2007 12:22 PM
To: tuscany-dev@ws.apache.org
Subject: A question of federation - was: Planning kernel release 2.0

On 3/22/07, Meeraj Kunnumpurath <[EMAIL PROTECTED]> wrote:
>
> Hi,
>
> Now that the SPI is getting stable and we have the initial end-to-end 
> story for federation working, I would suggest we plan for the final 
> release for kernel 2.0, with emphasis on federation and user
experience.
> I was thinking about aiming for a beta in June in time for TSSJS 
> Barcelona and the final release for August. Maybe we can have couple 
> of alpha releases from now and June as well. These are the features, I

> would like to see in 2.0.
>
> 1. Tidy up anything required in physical model, now that it is 
> starting to take good shape.
> 2. Tidy up generators from logical to physical model.
> 3. Fix the JXTA discovery issues, also investigate other discovery 
> protocols.
> 4. Federation end-to-end fully completed, this would include, maybe, 
> profiles advertising their capabilities and the information being used

> in intent-based autowiring etc.
> 5. Intent-based auto wiring
> 6. Emphasis on end user experience in terms of ease of use.
> 7. Assembly service, this kind of now related to the generators that 
> have been introduced in the last week or so 8. Artifact management, 
> especially mobile code when we target components to remote profiles.
>
> Also, now the SPI has started settling in, we need to start looking at

> binding and container extensions as well. Some of the bindings I would

> be interested in are,
>
> 1. JMS
> 2. AMQP
> 3. Hessian
>
> Ta
> Meeraj
>
>
> *
>
> You can find us at www.voca.com
>
> *
> This communication is confidential and intended for the exclusive use 
> of the addressee only. You should not disclose its contents to any 
> other person.
> If you are not the intended recipient please notify the sender named 
> above immediately.
>
> Registered in England, No 1023742,
> Registered Office: Voca Limited
> Drake House, Three Rivers Court,
> Homestead Road, Rickmansworth,
> Hertfordshire, WD3 1FX. United Kingdom
>
> VAT No. 226 6112 87
>
>
> This message has been checked for all email viruses by MessageLabs.
>
> -
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
> Hi Meeraj

From my perspective having demonstrable code in June would be spot on as
I have to speak on SCA then and would consider a demo if we could do it.

I don't have the knowledge yet to comment on the details of your
proposal just yet (hence the new subject) but a question. From a future
demo point of view I would like to show various runtime options some of
which are not federated  examples some of which are. Can I miss out the
federation bit if I want to? For example, I would potentially like to
show a variety of scenarios

- Hello world. the simplest possible single process example to get
people into how SCA works
- Standalone domain (a single VM)
 service provision (perhaps an AJAX style example where an SCA
composite provides services to the browser)
 service consumption (backend service access providing content to my
AJAX service)
- Federated domain (multiple VM)
 How SCA describes many connected composites.

I'm just starting now to look at how all the kernel stuff works so I
expect all this will become clear soon enough (I found your previous
posts giving explantion b.t.w - so am starting from there)

Regards

Simon


This message has been checked for all email viruses by MessageLabs.

This message has been checked for all email viruses by MessageLabs.

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



A question of federation - was: Planning kernel release 2.0

2007-03-22 Thread Simon Laws

On 3/22/07, Meeraj Kunnumpurath <[EMAIL PROTECTED]> wrote:


Hi,

Now that the SPI is getting stable and we have the initial end-to-end
story for federation working, I would suggest we plan for the final
release for kernel 2.0, with emphasis on federation and user experience.
I was thinking about aiming for a beta in June in time for TSSJS
Barcelona and the final release for August. Maybe we can have couple of
alpha releases from now and June as well. These are the features, I
would like to see in 2.0.

1. Tidy up anything required in physical model, now that it is starting
to take good shape.
2. Tidy up generators from logical to physical model.
3. Fix the JXTA discovery issues, also investigate other discovery
protocols.
4. Federation end-to-end fully completed, this would include, maybe,
profiles advertising their capabilities and the information being used
in intent-based autowiring etc.
5. Intent-based auto wiring
6. Emphasis on end user experience in terms of ease of use.
7. Assembly service, this kind of now related to the generators that
have been introduced in the last week or so
8. Artifact management, especially mobile code when we target components
to remote profiles.

Also, now the SPI has started settling in, we need to start looking at
binding and container extensions as well. Some of the bindings I would
be interested in are,

1. JMS
2. AMQP
3. Hessian

Ta
Meeraj


*

You can find us at www.voca.com

*
This communication is confidential and intended for
the exclusive use of the addressee only. You should
not disclose its contents to any other person.
If you are not the intended recipient please notify
the sender named above immediately.

Registered in England, No 1023742,
Registered Office: Voca Limited
Drake House, Three Rivers Court,
Homestead Road, Rickmansworth,
Hertfordshire, WD3 1FX. United Kingdom

VAT No. 226 6112 87


This message has been checked for all email viruses by MessageLabs.

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Hi Meeraj



From my perspective having demonstrable code in June would be spot on as I

have to speak on SCA then and would consider a demo if we could do it.

I don't have the knowledge yet to comment on the details of your proposal
just yet (hence the new subject) but a question. From a future demo point of
view I would like to show various runtime options some of which are not
federated  examples some of which are. Can I miss out the federation bit if
I want to? For example, I would potentially like to show a variety of
scenarios

- Hello world. the simplest possible single process example to get people
into how SCA works
- Standalone domain (a single VM)
service provision (perhaps an AJAX style example where an SCA composite
provides services to the browser)
service consumption (backend service access providing content to my
AJAX service)
- Federated domain (multiple VM)
How SCA describes many connected composites.

I'm just starting now to look at how all the kernel stuff works so I expect
all this will become clear soon enough (I found your previous posts giving
explantion b.t.w - so am starting from there)

Regards

Simon