added RideSharingDesign design solution

Author: Iurii-Dziuban

README.md

@@ -145,7 +145,7 @@ The platform currently supports a total of 11 languages: C, C++, Java, Python, C
   - `ParkingLotDesign` - design of Parking Lot
   - `WhatsAppDesign` - design of messaging system like whatsApp / facebook messenger /weChat
   - `TwitterDesign` - design of messaging system with ability to follow and have timeline
-
+  - `RideSharingDesign` - design of car sharing system.
 # Codility.com tasks
 Codility is a platform for hiring stronger programmers faster. Has requirements for time and space complexity.
 Look for "codility" package

src/test/java/iurii/job/interview/design/RideSharingDesign.java

@@ -0,0 +1,52 @@
+package iurii.job.interview.design;
+
+/**
+ * Ride sharing service like Uber or Lyft
+ * https://www.youtube.com/watch?v=J3DY3Te3A_A
+ * Overview:
+ * - Payment management
+ * - Customer feedback gathering
+ * Core features
+ * - Finding and matching customers to drivers
+ * - Mapping and routing ETA (estimated time of arrival)
+ *
+ * Bottom Up Approach - Storage - All trips - stores trips in multiple data centers and store to closest geo-located
+ * If offline one data center, store to another. Availability is important. Eventual consistency.
+ *
+ * Low latency calls are needed. Uber - sql and non sql. Cassandra / MySQL
+ * Store archive in Data Warehouse (Hadoop) to run more expensive queries. Business analytics
+ *
+ * Caching layer - in-memory database, to use for static data like maps, user data, but drivers positions are dynamic.
+ *
+ * Logging - which driver at which position, did what. As real time as possible.
+ * From endpoints : customers or drivers: small messages about positions.
+ * Cluster of Kafka servers is used to store the messages, ensure they are not lost and can be stream processed in real time
+ * Kafka can use Hadoop to store/persist data.
+ *
+ * SOA - service oriented architecture
+ * Provisioning services running - get software to newly running servers or containers. Ex. Terraform/Puppet/Docker - shielded containers,
+ * that feels like separate host, but in fact many docker containers are running on the same linux machine
+ *
+ * Mesos/Kubernetes - container orchestration tool:
+ * manages/coordinates collaboration between services. Manages life cycles of the services during runtime
+ *
+ * In order to connect client to particular service - network routing problem as efficient as possible
+ * and routing mechanism should know the state of the services. Because such routing is difficult -
+ * many services are designed stateless, so that it will not matter to which service client will be connected
+ *
+ * Testing:
+ * - Unit test
+ * - Distributed system with services linked and chained - resilience testing, in case part of system is not working.
+ * Tool Hailstorm - goes to some services and puts them offline on some of group of hosts (shadow fleets),
+ * that are on production but does not serve the customers. Log and observe, how system works.
+ *
+ * Graph problem (routing) city map to graph:
+ * - route from driver to you
+ * - route to destination
+ * Shortest way from one node to another: speed allowance - weighted edges (can be precomputed)
+ * ETA in real time (traffic information). Relying on historical data : past similar situations, fallback approach:
+ * NP - complete (Dejkstra or A*) to calculate ETA with that - split city into smaller parts that are independent (not easy)
+ * as possible and small graphs can be computed quickly.
+ */
+public interface RideSharingDesign {
+}