Ashok,

Replies are inline below.

Assuming that I really understand what you are proposing, I was interested
in this idea. However, after thinking about this over the weekend, I think
that this is a really bad idea. Here are the big issues that I have:

1)      The design is a violation of OIC architecture and design approach.
It is not RESTful.

2)      The interactions are clearly procedural behaviors.

3)      The payload is not a supported format. It is not JSON (or CBOR).

4)      The abstraction has issues as any client MUST understand Zigbee or
Z-wave commands so standard OIC clients cannot use this feature.

5)      The design, as presented, only provides access to local clients to
local hardware and most Zigbee and Z-wave developers would prefer to work
directly with the hardware SDK APIs.

We have a need for Zigbee on v1.0. Based on your design, this will not
fulfill our requirements. Any of the above is reason to defer this until the
issues are addressed. However, more importantly, we have precious little
time to make Iotivity product worthy. So, I must prioritize features that
someone wants to productize. 

Pat



From: [email protected]
[mailto:iotivity-dev-bounces at lists.iotivity.org] On Behalf Of ASHOKBABU
CHANNA
Sent: Sunday, March 22, 2015 10:56 PM
To: Keany, Bernie; Macieira, Thiago; iotivity-dev at lists.iotivity.org
Subject: Re: [dev] About ZigBee, Z-Wave Supporting



Hi Bernie,

There multiple ways to looks application developer's effort here :


1) New developers adapting OIC APIs:
      A) For ZigBee developer, it's easy to adapt if we provide OIC APIs
using encapsulation approach.
      B) For OIC developer, it's not easy because they have to understand
ZigBee specific commands but this is same in PPM approach as well.



[PCL, 2015/03/23] I do not think that this is true. An OIC developer does
not need to understand Zigbee at all using the PPM. They only need to
understand the OIC profile. When they say "turn on the light" as a OIC
message, it is the protocol translators' responsibility to understand how to
send the message as Zigbee. In this world, a light is a light regardless of
whether it is a Hue light bulb or a Zigbee lightbulb.



2)  Existing developers adapting OIC APIs:
     ---> Having PPM approach,  ZigBee/Z-Wave app developers need to
understand how OIC translate the ZigBee commands and also how the plugin
need to be installed for multiple ZigBee profile in the current market
scenario even though developer interested in particular ZigBee profile with
specific vendor chipset.
        so app developers effort is much more than second approach.



[PCL, 2015/03/23] The goal of OIC is not support Zigbee or Z-wave developers
in this way.



3) In both cases target is same.  
             OIC devices interact with non-OIC standard devices.


4) As there are multiple players in ZigBee and each have their distinction
of APIs, it might be difficult to select which plugin is suitable by the
developer.



I personally feel, connectivity's related to radio
(WIFI/BT/BLE/Zibgee/Z-Wave/IPv6  ...etc) need to be handled via CA and
protocol translations such as AllSeen,UPnP.. etc need to be handled via PPM.


There might be Architectural issues but unified approach will help in the
long-term for OIC.

And having additional benefits such as used of all OIC devices ( Rich/Lite),
new use-cases gives second approach a better option. 


Regards,
Ashok

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

Sender : Keany, Bernie<bernie.keany at intel.com
<mailto:bernie.keany at intel.com> >

Date : Mar 20, 2015 17:01 (GMT+09:00)

Title : Re: [dev] About ZigBee, Z-Wave Supporting



Hello Ashok

Your final comment to Thiago about increasing the effort on application
developers was a point I wanted to make based on my reading of Jinguk's
slides (slide 4).

I had always believed we were developing the APIs and so that application
developers didn't have to know about transports or physical layers. I
believe that this goal should be top of mind when it comes to architecture
and design decisions; so for that reason I feel the PPM approach is
preferred.

Regards,
Bernie

From: ASHOKBABU CHANNA <[email protected]
<mailto:ashok.channa at samsung.com> >
Reply-To: "ashok.channa at samsung.com <mailto:ashok.channa at samsung.com> "
<ashok.channa at samsung.com <mailto:ashok.channa at samsung.com> >
Date: Thursday, March 19, 2015 at 7:38 PM
To: "Macieira, Thiago" <thiago.macieira at intel.com
<mailto:thiago.macieira at intel.com> >, "iotivity-dev at lists.iotivity.org
<mailto:iotivity-dev at lists.iotivity.org> " <iotivity-dev at 
lists.iotivity.org
<mailto:iotivity-dev at lists.iotivity.org> >
Subject: Re: [dev] About ZigBee, Z-Wave Supporting


Hi,

Even though PPM approach is default approach to be considered,

Encapsulation (2nd approach) provides additional benefits such as



1) Used by all OIC devices (Rich/Light) .

2) Make unified approach for OIC devices like BT/BLE connectivity's.

3) New use-cases such as handling Zigbee devices at single hop or multi hop
with OIC framework.

        For ex::  If we have 3 devices - A ( OIC Device- BLE) , B ( OIC
Device - BLE/Zigbee) , C ( Zigbee device).

                    From A-B , it can be a OIC protocol(payload zigbee
command) and from B to C its Zigbee interaction using OIC framework itself.



And benefits will come with a cost.. it makes applicationput more effort.

But this is valid scenario as application have more idea of Zigbee Profile .



In my opinion, encapsulation is also good approach to be considered for
better OIC device expansion.



Regards,
Ashok

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

Sender : Thiago Macieira<thiago.macieira at intel.com
<mailto:thiago.macieira at intel.com> >

Date : Mar 19, 2015 23:13 (GMT+09:00)

Title : Re: [dev] About ZigBee, Z-Wave Supporting



On Thursday 19 March 2015 06:21:56 JinGuk Jeong wrote:
> Dear IoTivity Developers,
>
> Since I want to technically discuss about a way how to support
> ZigBee/Z-Wave, I am sending an email to you. I think this feature is very
> important feature because there are so many ZigBee/Z-Wave devices in the
> world. However, we need to consider several issues such as chip
dependency,
> code size, and so on. Currently, there are two approaches for that, and
you
> can find these approaches in attached file. Plz give your feedback about
> this.

Hi Jinguk

Thanks for sharing the outlines of the two options.

I'm probably missing something... it seems quite straightforward to me that
the simplest answer would be #1, which is why I think I am missing
something.
Why would option #2 be considered?

My thinking is:
1) the Zigbee/ZWave wire protocol and message format is wildly different
from
the OIC one (discovery, communication, authentication, encryption,
authorisation, etc.). That means the only abstraction that would be common
between OIC and a Zigbee device is at the level of a service. That is, a
lightbulb is still a lightbulb by any other name.

2) there are probably existing libraries to talk to zigbee adapters and
speak
zigbee protocols (or zwave), which means we would spend less time trying to
abstract them on CA. Instead, we gain time by simply wrapping them at the
highest level possible -- the service layer.

The only thing that comes to mind about wanting to have CA service zigbee is
if we wanted to share the infrastructure between multiple zigbee-capable
devices. But isn't that what the zigbee libraries should already be doing?
Do
such iibraries exist?

--
Thiago Macieira - thiago.macieira (AT) intel.com
  Software Architect - Intel Open Source Technology Center

_______________________________________________
iotivity-dev mailing list
iotivity-dev at lists.iotivity.org <mailto:iotivity-dev at lists.iotivity.org> 
https://lists.iotivity.org/mailman/listinfo/iotivity-dev





[http://www.samsung.net/service/ml/AttachController/201503201138460_QZUWXYH6
.gif?cmd=downdirectly&filepath=/LOCAL/ML/CACHE/a/20150323/93BIV0UVOUUB at namo.
co.kr705ashok.channa&contentType=IMAGE/GIF;charset=Windows-1252&msgno=705&pa
rtno=1&foldername=INBOX&msguid=71225]











</thiago.macieira at intel.com</iotivity-dev at 
lists.iotivity.org</thiago.macieir
a at intel.com</ashok.channa at samsung.com</ashok.channa at samsung.com


<http://ext.samsung.net/mailcheck/SeenTimeChecker?do=a89e2f31c590267a8e61892
067e2c10997670b70b8e6dc7ea1889aa8efbbee4713d69ea60670f66e1fac5bf6c61543587bc
8a6096cfbfc9f02e038c8d853bffbdb9fdddda33e82cbe4a391424e62fcf6cf878f9a26ce15a
0> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
<http://lists.iotivity.org/pipermail/iotivity-dev/attachments/20150323/e37ae4b4/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.gif
Type: image/gif
Size: 13168 bytes
Desc: not available
URL: 
<http://lists.iotivity.org/pipermail/iotivity-dev/attachments/20150323/e37ae4b4/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 7198 bytes
Desc: not available
URL: 
<http://lists.iotivity.org/pipermail/iotivity-dev/attachments/20150323/e37ae4b4/attachment.p7s>

Reply via email to