Comments on: Review of Fujitsu’s IaaS Cloud API submission to DMTF http://stage.vambenepe.com/archives/1108 IT management in a changing IT world Tue, 09 Mar 2010 22:10:17 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: William Vambenepe — Can Cloud standards be saved? http://stage.vambenepe.com/archives/1108#comment-102784 William Vambenepe — Can Cloud standards be saved? Mon, 15 Feb 2010 16:24:19 +0000 http://stage.vambenepe.com/?p=1108#comment-102784 [...] keeps the submission confidential. That’s why I blogged about the vCloud submission and the Fujitsu submission but couldn’t provide equivalent analysis for the [...] [...] keeps the submission confidential. That’s why I blogged about the vCloud submission and the Fujitsu submission but couldn’t provide equivalent analysis for the [...]

]]> By: The Fujitsu Cloud API http://stage.vambenepe.com/archives/1108#comment-100750 The Fujitsu Cloud API Mon, 21 Dec 2009 16:09:51 +0000 http://stage.vambenepe.com/?p=1108#comment-100750 [...] with time to blog so let me give you a quick run down of what’s going on and then link you to a great post by William Vanbenepe for more detailed [...] [...] with time to blog so let me give you a quick run down of what’s going on and then link you to a great post by William Vanbenepe for more detailed [...]

]]> By: William (@vambenepe on Twitter) http://stage.vambenepe.com/archives/1108#comment-100682 William (@vambenepe on Twitter) Sat, 19 Dec 2009 00:02:33 +0000 http://stage.vambenepe.com/?p=1108#comment-100682 Thanks Roger. BTW, I didn't mean to suggest that providing a good RESTful design is "a simple engineering task". What is a simple engineering task is to come up with something that looks RESTful on the surface. That's why I said "RESTful-looking (but contorted)". And that's what I see a lot of in "RESTful" Cloud APIs. Thanks Roger. BTW, I didn’t mean to suggest that providing a good RESTful design is “a simple engineering task”.

What is a simple engineering task is to come up with something that looks RESTful on the surface. That’s why I said “RESTful-looking (but contorted)”. And that’s what I see a lot of in “RESTful” Cloud APIs.

]]> By: Roger Menday http://stage.vambenepe.com/archives/1108#comment-100678 Roger Menday Fri, 18 Dec 2009 22:13:50 +0000 http://stage.vambenepe.com/?p=1108#comment-100678 William, I think the review is very fair,and, in particular your discussion around 'getVSYSDescriptor' is thought provoking. I do disagree with your assertion (backed by Tom) that producing a RESTful rendering is a simple engineering task. I think if that was the case, we would've seen this in the IaaS space already. Some are good, but, nothing is great - yet ! William, I think the review is very fair,and, in particular your discussion around ‘getVSYSDescriptor’ is thought provoking. I do disagree with your assertion (backed by Tom) that producing a RESTful rendering is a simple engineering task. I think if that was the case, we would’ve seen this in the IaaS space already. Some are good, but, nothing is great – yet !

]]> By: Tom Maguire http://stage.vambenepe.com/archives/1108#comment-99794 Tom Maguire Wed, 25 Nov 2009 19:00:49 +0000 http://stage.vambenepe.com/?p=1108#comment-99794 vbp, I couldn't agree with you more on some of your points: "“RESTful” rather than RPC approach" Personally, I prefer the noun centric approach (what's the resource model) to the verb centric approach (API/RPC) mostly because it forces us to talk about the abstractions that are being managed. I think that rather than esquisse (love that word by the way) I would say it is inchoate. You are right to point out however that mapping between an XML/RPC style and a RESTful style is more an exercise for the engineer than a really challenging problem. BTW, that mapping forces us to discuss the abstractions which is a good thing. On a related point, the application layer protocol wars used to upset me but now just cause me unbearable ennui. “model-based versus procedural” debate I lean much more toward the declarative model-based approach. We've been doing the procedural deal forever and it has well known issues and constraints. That being said, I believe procedural has its place, just not in the cloud. "(VSYS Descriptor) intends to build on OVF" Amen, the addition of requirements to the OVF is the next big thing that needs to be done. Those requirement expcession ill allow for a loosely coupled match between the requirements and the capabilities of the infrastructure. All good stuff, I look forward to the DMTF's work in this area. Tom vbp,
I couldn’t agree with you more on some of your points:

““RESTful” rather than RPC approach”

Personally, I prefer the noun centric approach (what’s the resource model) to the verb centric approach (API/RPC) mostly because it forces us to talk about the abstractions that are being managed. I think that rather than esquisse (love that word by the way) I would say it is inchoate. You are right to point out however that mapping between an XML/RPC style and a RESTful style is more an exercise for the engineer than a really challenging problem. BTW, that mapping forces us to discuss the abstractions which is a good thing. On a related point, the application layer protocol wars used to upset me but now just cause me unbearable ennui.

“model-based versus procedural” debate

I lean much more toward the declarative model-based approach. We’ve been doing the procedural deal forever and it has well known issues and constraints. That being said, I believe procedural has its place, just not in the cloud.

“(VSYS Descriptor) intends to build on OVF”

Amen, the addition of requirements to the OVF is the next big thing that needs to be done. Those requirement expcession ill allow for a loosely coupled match between the requirements and the capabilities of the infrastructure. All good stuff, I look forward to the DMTF’s work in this area.

Tom

]]>