Following are questions people asked during talks given on the EASM
protocol. 


Questions at Honeywell, Bangalore, India
------------------------------------------

(1) priority handling and dissemination of messages
(2) overhead of initial setup of applying hash function.
(3) how is the hash function known to all parties
(4) what happens when hash function results in lots of collisions due
    to which rendezvous point gets overloaded
(5) load balancing at rendezvous point? If done what is the messaging
    overhead
(6) how is security handled?
(7) how is refreshing of data in rendezvous node done i.e., say
    sources may die or phenomenon may move, so we need to modify info
    stored in rendezvous nodes.
(8) what if a sink has moved far away from the rendezvous point - in
    that case can we come up with better hash functions that will
    realize rendezvous points that are optimally placed from the sink
    as well as the source?
(9) what if the source has to send the same data to multiple sinks -
    are we doing some optimizations w.r.t energy savings e.g., if two
    sinks are in the same cell, do we send the message to the
    grid center only once?
(10) Say the rendezvous node has relayed a request for data type XYZ
     from some sink to a source at time, t. Now say some another
     sink asks for the same data type at time t + delta. If these
     requests are so close to each other, do we provide any
     optimizations to reduce messaging traffic.
(11) What if some nodes in the network decide to sleep. What is the
     effect of that on GPSR?
(12) what if the source dies - how is the sink notified?
(13) can many sources cooperate on the same data collection (I think
     this had to do with data aggregation)

Honeywell people had slightly different requirements in terms of their
sensor networks. For them power is crucial. So they will also try to
do what they call power scavenging.

Discussions at Veritas
----------------------
(1) How is state transferred from original rendezvous node to
    replicas. What is the effect on hash function? Can hash functions
    change on the fly so they can point to the new rendezvous node?
(2) what if a rendezvous node suddenly dies and not due to battery
    going below threshold? How does our fault tolerance scheme work.
(3) When we say that a sink is interested in a data type, are we
    describing an instance of the data type or any instance that
    belongs to that class.
(4) To address the "t+delta" issue raised by Honeywell, Bala suggested
    using a time-to-live until which the rendezvous node waits to
    relay the message. 
(5) why cannot normal handoff schemes from wireless networks be used
    in WSNs?
(6) how is the hash fn deployed in the nodes? Can it be dynamically
    configured?
(7) What is the relationship to RFIDs? Is it mirror image?
(8) One graph talked about failures? What does 98% success mean? Is
    this a function of overall failure distribution?
(9) How did we measure energy (in Joules) ?