In the latest version of Android, Google shifted from bitmaps to vector rendering for maps. The shift is probably more a function of available processing power on the device than a strategic shift to position value on the device rather than in the cloud. Vector maps, which are much more efficient in terms of bandwidth and local storage, have been the choice for in-car navigators but Google has always been using bitmap tiles which are fetched from a server and delivered only if the device is using a data connection. The downside for vectors is that they require a bit more local processing power.
I doubt that Google’s move to vectors is part of a shift to more app-centered/edge-of-network strategy.
How is this relevant?
Google has been riding a wave of re-centralization of value toward the center of the network as broadband made the cloud feasible. Keep in mind that Microsoft rode a wave of de-centralization as value moved to the PC and away from the mainframe. As intelligence was pushed to the edge, Microsoft accrued value from enabling the edge as a locus of productivity. This is no small thing.
Then came apps.