![]() ![]() If the data changes further down the line, change the schema and you are safe. You can define permitted and expected kinds of relationships and provide patterns of anticipated data trees. Providing codified schemas to code so that it can explore data trees in a meaningful manner (removing guess work) is a helpful approach. I solved some of these issues with meta-data tags, recursive algorithms and an interlingua I call dTree. To add to the fun I did it with a typed language! Another gotcha came in the form of preserving ordered data when transorming from a representation such as JSON or XML to a graph and back again. When building out the prototypes for exploring GraphOps, one of the biggest issues was writing code that dealt with arbitrary graph structures and relationships. Usenix MALT paper link Some Notes on Writing Code MALT appears to be reasonably useful and more importantly, used! The paper is an interesting read and there is lots of common sense shared with readers. What I saw when the film hit the cinema excited me. Glimpse into the imagination of someone thinking about SkyNet.She’s software savvy and opened me up to levels of depths I didn’t know existed. ![]() A fantastic mathematician and available for work (if you didn’t know that!).Rachel Traylor’s education on Graph Theory and Reliability (see IPSpace offerings) John Ousterhout - A Philosophy of Software Design Mark Burgess and Jan A.Bergstra - Promise TheoryĮdsger W.Dijkstra - A Discipline of Programming NANOG81: Navigating Automata Without a Treasure Map Inspiration IPSpace Autumn 2020: Evolution of Network Automation Imagine knowing the what, where, why and how for your infrastructure and service delivering resources - always. I’m firmly of the belief the way forward for operations is in this manner. I’ve been lucky that different groups have humoured me with mad ramblings. It becomes possible to make a change and verify or validate the outcome and even provides percepts for machine-learning and artificial intelligence. With a rich tapestry like a graph representation of your infrastructure, you have a foundation for self-driving and self-healing systems. Imagine a world where an unexpected issue triggers the generation of a blast radius linked tree, highlighting outcomes, failures and steps taken to recover. They are alive, not quite like Johnny 5, but in the sense of continually reflecting state and relationships. This isn’t staring at XY scatter plots, but dealing with infrastructure as a richly connected graph of resources that either consume or provide services to and from users. This page contains information I’ve been gathering in the name of an emerging branch of operations I’m dubbing “GraphOps”.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |