Having global LPM tables to classify source addresses

Issue #208 introduced LPM tables for IPv4 and IPv6 addresses to policies. The main use of these LPM tables is to classify source addresses with a single lookup. These LPM tables are instantiated and populated on all GT blocks with the same datasets. As the number of these datasets grow, the amount of duplicate memory grows by the product of the number of GT blocks. Therefore, if these tables are made global to all GT blocks, a lot of memory will be saved. Moreover, DPDK's LPM tables already support current lookups, so the effort to have global LPM tables available to policies is limited to Gatekeeper itself.