====== HW Visit 2016-10-27 ======

Continue SP's notes..

Discuss anomalies in the result of Reactive, proactive delays

ZD: HO delays depends on the location of MH, it is the time b/w successful send pkt at origin and target AP (???)

ZD: b/w Mh at AP1 has pkt to send until it can send at AP2 ( Max(flows are ready, L2 delay) ). If flow is not ready, the sent pkt will be pk_IN to Ctrl.

MK: in the downlink the time b/w rcv pkt at orgin and rcv pkt at target. 

MK: We agree to use ZD's measurement suggestion.

ZD: We still see value of the testbed if we can't understand single detail of the result, e.g., L2 delay, total HO, ... (sell the testbed's physical measurement vs. mininet)

MK: has an idea to use TTT (learning) to shift Ctrl load before UE disassociate from origin eNB.

MK: postpone S2S until Tobi is back

Discuss deliverable

MK: we combine D7 D8 and provide testbed documentation instead of D7

SP: Scope changed: we moved away from cloud computing.

ZD: OK for me. Otherwise we put 3GPP IPR in D7 and keep them 7, 8 separated.

MK: Scenario