i.e. expect all services of RackHD run on same machine or not
i.e. we need to scale beyond 1 rack, or 1 row
Former user (Deleted): neighborhoods that span racks, rows, DCs and geographic sites
Former user (Deleted): Re-confirm the notion of a single instance of "RackHD Core" service are responsible for a "neighborhood", no matter how large or small
Former user (Deleted): Must make some statements about how large a neighborhood can be
Former user (Deleted): We have to express requirements in terms of SLA? (what metrics do we express our scale requirements in?)
No. of nodes, - not enough to express in this term
No. of workflows / sec
What about discovery?
What about OS install?
What about firmware updates?
Former user (Deleted): Need to be able to say that we have architect-ed this to address the different aspects of scale
Former user (Deleted): How do we know when we have broken down the services to a "granular enough" level – what is practical? How do we know what is practical?
Former user (Deleted): Need to be cognizant of unintended consequences that come from granularity
Discuss 1st deliverable - updated architecture diagram – architecture as it is today
What services in ecosystem should be included as being part of the architecture?
Should we include all services, no matter how mature?
Former user (Deleted): AI: Create strawman set of diagrams that include all RackHD Core, some external, some InfraEnablers and Enhanced PE
Deployment diagrams vs. architecture diagrams
Solution architecture vs. design architecture
What should an architecture diagram include?
How much detail is appropriate?
Should we have a "drill down" approach to architecture diagrams?
Former user (Deleted): Agrees that more detali is needed for some of our discussions
Former user (Deleted): We need to have a single source of truth for what this team feels is "our architecture" that others can look to for builiding their own collateral for their own audiences
Tim Larson: AI: Joe and Tim to pull together existing collateral into a single place
Review AIs from 12/19/16 OSE Commit Review Board meeting
Former user (Deleted): Did not get to this set of topics, will need to be added to the backlog of topics
Deeper discussion(s) on the following topics
Image repository service
Service discovery and the use (or not) of Consul
Service isolation & decomposition
How granular is practical?
What is driving the granularity?
Action items
Former user (Deleted): - Strawman mission/vision statement(s) - posted in Works In Progress
Former user (Deleted), Tim Larson: - Pull together existing architectural collateral into a single place - Worked with Tim L to gather various decks from his working space along with published decks from Syncplicity
Former user (Deleted): - Strawman (starting set) architectural diagrams of where we are today (drill downs) - have not generated a strawman, but have collected what I believe to be the set of collateral that represents our latest diagrams – these will be used for discussion and how to move forward creating our new single source of truth updated to today.