[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
RE: [hyades-dev] HCE Protocol Layer 2: Agent Interface


My response to Allan's comments addresses some of the concerns
raised by you.  Here are some more.

Get Agent ID - I am removing this one.

CID_AGENT_RESET - There could be a scenario where a client will configure and
use the agent for some time and leave it idle before using it again with different configuration.
Or it could be another client that will start to use an existing agent. So, it could be easy
if we reset the configuration of the agent.

the whole query interface as these are already available in agent manager interface (as you mentioned).

CID_COLLECTOR_EVENT_PROPRIETARY - We should define how to use these commands (let us review tomorrow)

Thanks for taking time reviewing the document and sending your comments.

Intel Corporation

> -----Original Message-----
> From: hyades-dev-admin@xxxxxxxxxxx
> [mailto:hyades-dev-admin@xxxxxxxxxxx]On Behalf Of Plaetinck, Luc
> Sent: Wednesday, August 11, 2004 9:22 AM
> To: 'hyades-dev@xxxxxxxxxxx'
> Subject: [hyades-dev] HCE Protocol Layer 2: Agent Interface
> Some comments in addition to Allan's comments.
> General:
> What I would like to see also is a set of APIs (methods) for the
> HyadesBaseAgent which are necessary for all derived agents, 
> for example
> initialize(), how to add command listeners, client listeners and so on
> Specific:
> CID_AGENT_INIT: why is this necessary? Isn't it the 
> task of the
> agent at startup to initialize itself (which includes also registering
> itself with the HCE). If indeed some initialization is needed before
> performing a specific task like data collection, that should 
> be internal to
> the agent so a client doesn't have to worry about it: client 
> should only
> have to issue the command START_COLLECTING, and the agent 
> upon receipt of
> the command initializes/resets itself to start collecting
> CID_AGENT_DEREGISTER: is it the task of the client 
> (or HCE) to
> deregister an agent or should the agent deregister itself 
> when done and
> terminated. See command CID_DEREGISTER_AGENT in the Hyades Protocol
> Specification document
> Get Agent ID: if only for the agent's internal use, this is
> unnecessary. Agent ID is returned by the HCE through the command
> CONNECTION_COMPLETE (in Connection ID) when an agent first 
> connects to the
> HCE and the connection is succesful.
> CID_AGENT_RESET: personally I think this should not be the
> client's task. If an agent needs to be reset to initialized 
> state, that
> should be an internal task performed before every operation 
> that requires
> it. Other than that, a client should only have to issue 
> start/stop commands
> for a given task and not being bothered with agent cleanup
> CID_AGENT_UN_INITIALIZE: don't see the need for 
> this. See remarks
> for and
> CID_AGENT_GET_INTERFACES: wouldn't this be better 
> as a command for
> the HCE (see CID_GET_AGENT_METADATA). As all communication 
> goes through the
> HCE anyway, and the HCE reads the agent registry, there 
> really is no need
> for this command on an agent level. Furthermore, does it make 
> sense to have
> to attach to an agent in order to obtain info about what 
> interfaces the
> agent support?
> CID_AGENT_INTERFACES_SUPPORTED: see previous remark, also
> partially available through CID_QUERY_AVAILABLE_AGENTS 
> this works OK for the events
> described in, as these are predefined. However, as 
> Allan pointed
> out, there could be agent-specific events. An additional 
> command would be
> would not specify
> any event interface IDs.
> with Bytes 0-3 a Proprietary Event Identifier
> Luc
> The contents of this e-mail are intended for the named 
> addressee only. It
> contains information that may be confidential. Unless you are 
> the named
> addressee or an authorized designee, you may not copy or use 
> it, or disclose
> it to anyone else. If you received it in error please notify 
> us immediately
> and then destroy it. 
> _______________________________________________
> hyades-dev mailing list
> hyades-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/hyades-dev