1. Topics
  2. DRAFT slides for TOC meeting tomorrow

DRAFT slides for TOC meeting tomorrow

alexis richardson
 

Doug Davis <dug@...>
 

On the new reference architecture picture I have a few comments/questions:
1 - what is driving the need for a new one rather than tweaking the existing one?
2 - I actually prefer the old one. I like that it (abstractly) mentions most of the key pieces of the puzzle - e.g. the image registry and networking - things the new one doesn't touch on.

thanks
-Doug
_______________________________________________________
STSM | IBM Open Source, Cloud Architecture & Technology
(919) 254-6905 | IBM 444-6905 | dug@...
The more I'm around some people, the more I like my dog

Alexis Richardson via cncf-toc ---07/19/2016 04:42:16 PM---all, https://docs.google.com/presentation/d/1h8HFOAVLrJTvjUPP6ZHG2SzVKsTaMLvbiqjzUj4vml8/edit?ts=578

From: Alexis Richardson via cncf-toc <cncf-toc@...>
To: cncf-toc@...
Date: 07/19/2016 04:42 PM
Subject: [cncf-toc] DRAFT slides for TOC meeting tomorrow
Sent by: cncf-toc-bounces@...




all,

https://docs.google.com/presentation/d/1h8HFOAVLrJTvjUPP6ZHG2SzVKsTaMLvbiqjzUj4vml8/edit?ts=578d8da0#slide=id.g15e23e5137_1_0

comments actively sought

a
_______________________________________________
cncf-toc mailing list
cncf-toc@...
https://lists.cncf.io/mailman/listinfo/cncf-toc
alexis richardson
 

Doug,


On Wed, Jul 20, 2016 at 12:57 PM, Doug Davis <dug@...> wrote:

On the new reference architecture picture I have a few comments/questions:
1 - what is driving the need for a new one rather than tweaking the existing one?

IMHO the old one is quite hard to understand, especially for developers who might use CNCF projects.  

This is an attempt to simplify the picture somewhat.  At the end of the day most developers care about CNCF's main concerns: containers, runtime platforms, core services and ancillary tooling, and how they relate to each other, and to apps in general.


 

2 - I actually prefer the old one. I like that it (abstractly) mentions most of the key pieces of the puzzle - e.g. the image registry and networking - things the new one doesn't touch on.

I think the old one is really abstract -- much too abstract -- and contains extra info, eg about all the interop surfaces, which are unlikely to be seen or used by many developers.

Networking *is* in the stack -- it's part of the runtime layer.  This is covered in the "detail" presentation of the stack that Ken showed two weeks ago.

Registries are part of app dev.

a




 


thanks
-Doug
_______________________________________________________
STSM | IBM Open Source, Cloud Architecture & Technology
(919) 254-6905 | IBM 444-6905 | dug@...
The more I'm around some people, the more I like my dog

Alexis Richardson via cncf-toc ---07/19/2016 04:42:16 PM---all, https://docs.google.com/presentation/d/1h8HFOAVLrJTvjUPP6ZHG2SzVKsTaMLvbiqjzUj4vml8/edit?ts=578

From: Alexis Richardson via cncf-toc <cncf-toc@...>
To: cncf-toc@...
Date: 07/19/2016 04:42 PM
Subject: [cncf-toc] DRAFT slides for TOC meeting tomorrow
Sent by: cncf-toc-bounces@...




all,

https://docs.google.com/presentation/d/1h8HFOAVLrJTvjUPP6ZHG2SzVKsTaMLvbiqjzUj4vml8/edit?ts=578d8da0#slide=id.g15e23e5137_1_0

comments actively sought

a
_______________________________________________
cncf-toc mailing list
cncf-toc@...
https://lists.cncf.io/mailman/listinfo/cncf-toc
Doug Davis <dug@...>
 

hmm I guess it depends on the audience then. I might agree that from an app developer perspective the points of interop in the old pict might not be of great interest, but from a CNCF project perspective, and as someone working on the projects that make up the CNCF family of projects, I actually think that might be critical as that shows where in the overall picture we might be missing some community focus on projects in that space. Perhaps we need both?

thanks
-Doug
_______________________________________________________
STSM | IBM Open Source, Cloud Architecture & Technology
(919) 254-6905 | IBM 444-6905 | dug@...
The more I'm around some people, the more I like my dog

Alexis Richardson ---07/20/2016 08:11:32 AM---Doug, On Wed, Jul 20, 2016 at 12:57 PM, Doug Davis <dug@...> wrote:

From: Alexis Richardson <alexis@...>
To: Doug Davis/Raleigh/IBM@IBMUS
Cc: cncf-toc@...
Date: 07/20/2016 08:11 AM
Subject: Re: [cncf-toc] DRAFT slides for TOC meeting tomorrow




Doug,


On Wed, Jul 20, 2016 at 12:57 PM, Doug Davis <dug@...> wrote:
    On the new reference architecture picture I have a few comments/questions:
    1 - what is driving the need for a new one rather than tweaking the existing one?

IMHO the old one is quite hard to understand, especially for developers who might use CNCF projects.  

This is an attempt to simplify the picture somewhat.  At the end of the day most developers care about CNCF's main concerns: containers, runtime platforms, core services and ancillary tooling, and how they relate to each other, and to apps in general.


 
    2 - I actually prefer the old one. I like that it (abstractly) mentions most of the key pieces of the puzzle - e.g. the image registry and networking - things the new one doesn't touch on.

I think the old one is really abstract -- much too abstract -- and contains extra info, eg about all the interop surfaces, which are unlikely to be seen or used by many developers.

Networking *is* in the stack -- it's part of the runtime layer.  This is covered in the "detail" presentation of the stack that Ken showed two weeks ago.

Registries are part of app dev.

a




 
alexis richardson
 



On Wed, Jul 20, 2016 at 1:20 PM, Doug Davis <dug@...> wrote:

hmm I guess it depends on the audience then.

Yes.

 

I might agree that from an app developer perspective the points of interop in the old pict might not be of great interest, but from a CNCF project perspective, and as someone working on the projects that make up the CNCF family of projects, I actually think that might be critical as that shows where in the overall picture we might be missing some community focus on projects in that space.

When Ken and I created this stack, we first analysed all the projects that we could think of, that are in the space.  I want to share this info ASAP, but need to spend an hour cleaning it up.

 

Perhaps we need both?

I think the stack marketecture is of limited value unless backed up by:

1 -- detailed breakdowns of each layer's concerns into subcategories (eg orchestrator); and insofar as it exists any internal structure (eg relating orchestrator to container).

2 -- example projects for each subcategory; eg. "kubernetes is an orchestrator", collectively forming a market landscape 

We'd love help mapping the 'old' stack concepts into this model, as a proving exercise.

 



thanks
-Doug
_______________________________________________________
STSM | IBM Open Source, Cloud Architecture & Technology
(919) 254-6905 | IBM 444-6905 | dug@...
The more I'm around some people, the more I like my dog

Alexis Richardson ---07/20/2016 08:11:32 AM---Doug, On Wed, Jul 20, 2016 at 12:57 PM, Doug Davis <dug@...> wrote:

From: Alexis Richardson <alexis@...>
To: Doug Davis/Raleigh/IBM@IBMUS
Cc: cncf-toc@...
Date: 07/20/2016 08:11 AM
Subject: Re: [cncf-toc] DRAFT slides for TOC meeting tomorrow




Doug,


On Wed, Jul 20, 2016 at 12:57 PM, Doug Davis <dug@...> wrote:
    On the new reference architecture picture I have a few comments/questions:
    1 - what is driving the need for a new one rather than tweaking the existing one?

IMHO the old one is quite hard to understand, especially for developers who might use CNCF projects.  

This is an attempt to simplify the picture somewhat.  At the end of the day most developers care about CNCF's main concerns: containers, runtime platforms, core services and ancillary tooling, and how they relate to each other, and to apps in general.


 
    2 - I actually prefer the old one. I like that it (abstractly) mentions most of the key pieces of the puzzle - e.g. the image registry and networking - things the new one doesn't touch on.

I think the old one is really abstract -- much too abstract -- and contains extra info, eg about all the interop surfaces, which are unlikely to be seen or used by many developers.

Networking *is* in the stack -- it's part of the runtime layer.  This is covered in the "detail" presentation of the stack that Ken showed two weeks ago.

Registries are part of app dev.

a




 
NASSAUR, DOUGLAS C <dn283x@...>
 

I'm almost done with draft one of the periodic table of cloud native elements for you guys to throw rocks at. We should then align with ref arch   

Regards, Doug

On Jul 20, 2016, at 7:29 AM, Alexis Richardson via cncf-toc <cncf-toc@...> wrote:



On Wed, Jul 20, 2016 at 1:20 PM, Doug Davis <dug@...> wrote:

hmm I guess it depends on the audience then.

Yes.

 

I might agree that from an app developer perspective the points of interop in the old pict might not be of great interest, but from a CNCF project perspective, and as someone working on the projects that make up the CNCF family of projects, I actually think that might be critical as that shows where in the overall picture we might be missing some community focus on projects in that space.

When Ken and I created this stack, we first analysed all the projects that we could think of, that are in the space.  I want to share this info ASAP, but need to spend an hour cleaning it up.

 

Perhaps we need both?

I think the stack marketecture is of limited value unless backed up by:

1 -- detailed breakdowns of each layer's concerns into subcategories (eg orchestrator); and insofar as it exists any internal structure (eg relating orchestrator to container).

2 -- example projects for each subcategory; eg. "kubernetes is an orchestrator", collectively forming a market landscape 

We'd love help mapping the 'old' stack concepts into this model, as a proving exercise.

 



thanks
-Doug
_______________________________________________________
STSM | IBM Open Source, Cloud Architecture & Technology
(919) 254-6905 | IBM 444-6905 | dug@...
The more I'm around some people, the more I like my dog

<graycol.gif>Alexis Richardson ---07/20/2016 08:11:32 AM---Doug, On Wed, Jul 20, 2016 at 12:57 PM, Doug Davis <dug@...> wrote:

From: Alexis Richardson <alexis@...>
To: Doug Davis/Raleigh/IBM@IBMUS
Cc: cncf-toc@...
Date: 07/20/2016 08:11 AM
Subject: Re: [cncf-toc] DRAFT slides for TOC meeting tomorrow




Doug,


On Wed, Jul 20, 2016 at 12:57 PM, Doug Davis <dug@...> wrote:
    On the new reference architecture picture I have a few comments/questions:
    1 - what is driving the need for a new one rather than tweaking the existing one?

IMHO the old one is quite hard to understand, especially for developers who might use CNCF projects.  

This is an attempt to simplify the picture somewhat.  At the end of the day most developers care about CNCF's main concerns: containers, runtime platforms, core services and ancillary tooling, and how they relate to each other, and to apps in general.


 
    2 - I actually prefer the old one. I like that it (abstractly) mentions most of the key pieces of the puzzle - e.g. the image registry and networking - things the new one doesn't touch on.

I think the old one is really abstract -- much too abstract -- and contains extra info, eg about all the interop surfaces, which are unlikely to be seen or used by many developers.

Networking *is* in the stack -- it's part of the runtime layer.  This is covered in the "detail" presentation of the stack that Ken showed two weeks ago.

Registries are part of app dev.

a




 



_______________________________________________
cncf-toc mailing list
cncf-toc@...
https://lists.cncf.io/mailman/listinfo/cncf-toc
1 - 6 of 6