And so it goes on! I have additional progress to report on the open-source-friendly ExperiaSphere model for management and orchestration for SDN, NFV, and the cloud. I’ve completed my preliminary webinar series with major stakeholders in the operator and standards space, and it’s taught me a lot.
One thing it’s taught me is that explaining this on a webinar, even at a high level, is very difficult. We’ve typically needed a hour and a half to simply introduce the model and to map it to open-source elements at a very high level. My detailed slide deck, what I’ve been using to develop a complete explanation to the level needed to drive implementation, is already north of 70 slides, more than double the size of the high-level deck that took an hour and a half. You can do the math.
What this means is that I am not going to schedule further webinars on the model. It’s simply not going to be possible to make them truly valuable. Instead what I plan to do is to work harder on a series of public tutorial videos and slide decks. I’m still planning to be able to release these in late July or early August, because the time I save in not doing other webinars can be directed at building the greater level of content detail I now know I’ll need. Please check this blog, my LinkedIn Group (ExperiaSphere), my Google+ Channel (TomNolle) for information on when the videos will be available. The good news is that you can play, stop, and start a video as many times as needed. Large-audience webinars clearly will not work.
I’ve also been able to assemble enough information to do the critical mappings between the model and open-source tools. In many cases I’ve identified alternate approaches that could be explored for situations where cost-benefit optimization is essential. Let me go through the major areas to summarize.
In the critical central service data model, I’ve defined the structure of a service as a hierarchical collection of subordinate components that build downward to the actual resources. I’ve determined how this model can be specified and how it is linked to deployment and lifecycle management. While the approach I’m proposing is very “cloud-like”, it is also compatible with SDN and even with legacy elements.
I’ve identified how to extend the NFV IGS’s notion of a “Virtual Infrastructure Manager” to support any link between a functional vision of a service and the corresponding structural representation of components. My “Infrastructure Manager” will codify both the deployment and management hand-off at the function/structure boundary, which means it identifies how you present and utilize all management data, including that for SDN.
In the management area, I’ve identified a management model that conforms to i2aex proxy repository principles and that provides for the collection of management data from all “real” sources and its dissemination in any logical, structured, form that’s required. There are a number of enhancements to this area that seem to have potential, including the use of complex event processing (CEP) and big data.
The agile projection of management data is an element in lifecycle management and operations integration, but the other element is also provided in the model. Service components at any level are finite-state machines that can define states and process events—some of which come from other components and some from underlying management repository processes. The intersect of the state and event defines the lifecycle process to be run at that point, and so NMS, OSS/BSS, and other processes are integrated into a cohesive lifecycle vision through a common coupling at the service component level.
What my tutorial video is going to do is explain all of this in enough detail that a software team could then go out, get the identified open-source pieces, and integrate them into a structure for open management and orchestration that would work for any resource set that can be represented through one of my “Infrastructure Managers”. I’m diddling different notions of how to structure multiple tutorials to avoid one of those “Gone With the Wind” endurance contests, viewing-wise.
I’m sure people will have questions on this stuff. Those clients who want to do consulting on it can obviously do that, but for the rest what I’ll have to do is devise an economically reasonable way of answering inquiries. What I’m doing there is using the LinkedIn ExperiaSphere main group. If somebody posts questions or comments there, I’ll do my best to respond. I’m not going to do phone or webinar work for this sort of support because I can’t absorb the loss of billable time. Thus, if you are interested in getting questions answered on the model, you’ll need to join the group. Nothing will be done there other than Q&A; nobody can post without moderation so there will be no spam or job offers. I’ll post status and answer questions, and that’s that.
If you’ve indicated interest in one of the webinars that were to be done in June or July, I’m sorry to say they won’t be happening. Not because I don’t want to, but because I know now from experience that the format won’t be enough. I’ll be putting video tutorials up as soon as possible, and I’m pretty confident that it will be worth the wait.