>> the taxonomy of this space
Aha!
1. Express the taxonomy of the internet resource universe as a C++ class tree ... abstractions ranging from pages, users, hits, caches ... down to servers ... down to routers, switches, and pipes.
2. Instance the class tree with actual data corresponding to the real-world web. Or, more generally, require internet "resources" to respond to a resource identification protocol ... current hot net design area. This would allow the web emulation to discover/learn resource/topology changes.
Once you have the object tree, and an instanced version of the object tree, you can develop various functions on it:
- EmulateTheWeb()
- ExtrapolateTheWebForward()
- IdentifyWebBottlenecks()
- ProposeMostUsefulUpgrades()
Instancing the object tree, and implementing the functions across disparate layers of heterogeneous objects is a complex problem.
Optimizing gates and cells on multi-vendor silicon is a problem of the same order? Whose solution offers insights into how the meta-web might be actually modeled?
Silicon vendors currently provide HDL and VHDL language descriptions of their components, which are sucked into design systems.
Network vendors could provide similar models of their components which could be compiled into the web object tree.
These models could include control/status info (ala SNMP) which would allow functions on the web model to interrogate and control it's physical corollary.
Thus, from the intersection of CAD design, emulation, network management software, etc ... possibly appears the synthesis of the meta-web ... one pathway.
So one identity of the meta-web is as a VVV-HLD model of the real web ... and useful functions performable on that model ... and the software technology which makes it possible.
Contrast this identity with that mentioned previously ... as a meta- operating system which learns the structure of the resources it manages.
Does this imply ... meta(1)-web, meta(2)-web ... at different levels of abstraction? |
