statusservice.html

m4_include(/mcs/m4/worksp.lib.m4)
_NIMBUS_HEADER(WSRF Interfaces)
_NIMBUS_HEADER2(n,n,y,n,n,n,n)
_NIMBUS_INTERFACES_WARNING
_NIMBUS_LEFT2_COLUMN
_NIMBUS_LEFT2_INTERFACES_SIDEBAR(n,n,n,n,n,n,y)
_NIMBUS_LEFT2_COLUMN_END
_NIMBUS_CENTER2_COLUMN
_NIMBUS_IS_DEPRECATED

<a name="service"></a>

<h2>Workspace Status Service</h2>

<p>
    The status service allows a client to query the usage data the service
    has collected about it.
</p>

<p>
    Currently there are two operations, <b>queryCurrentWorkspaces</b> and
    <b>queryUsedAndReserved</b>.
</p>

<hr />
<p>
    <b>queryCurrentWorkspaces</b> returns some status information on all
    instantiated workspaces that the caller controls.  It returns the schedule,
    network addresses, ID number, etc. for each.
</p>

<hr />
<p>
    <b>queryUsedAndReserved</b>
    returns the time (in minutes) used so far by the client as well as the
    currently reserved time.
</p>

<p>
    The <b>used time</b> is measured from creation time <i>to WSRF resource
    destruction time</i>.  Note that the end of the measurement is not
    when the workspace's running time is over.  Thus if you have lengthy
    stage-out transfers or extend the resource termination time for no
    good reason (which ties up allocations, including networking addresses),
    you will still be charged for this time.
</p>

<p>
    The <b>reserved time</b> is the aggregate of all your
    <i>current deployments'</i> requested durations.  When a deployment ends,
    the used time for that deployment is added to 'used' and the appropriate
    time is substracted from 'reserved'.
</p>

<p>  
    Reserved time is tracked separately mainly for authorization
    purposes: the authorization policy can look at your past and current 
    deployments together to render a decision.  If 'used' were the only thing
    tracked and consulted during authorization, a client could conceivably
    "blast" the service with requests that would ultimately put it above
    its quota down the line because the current used tally would be low at
    creation time.
</p>

<p>
    There is currently one resource property <b>chargeGranularity</b>.  This
    is an integer that tells how running time is being "counted".  If this
    value is one then anything between zero and one minute (inclusive) is
    counted as one minute.  If this value is e.g. thirty, then anything between
    zero and thirty minutes (inclusive) is counted as thirty minutes.
</p>

<p>
    Why would chargeGranularity ever be set differently than a value of one?
    There may be organizational accounting reasons, but the workspace service
    backend is usually the cause of this.  For example, the alternative backend
    to Amazon's Elastic Compute Cloud would cause the chargeGranularity to be
    sixty.  On this backend, the same amount of money is currently charged for
    any time used between zero and sixty minutes (inclusive), and so this
    parameter can be altered to accurately express a usage value that directly
    translates to the real charges.
</p>

<p>
    The Status Service WSDL can be
    <a href="../examples/compact/workspace_status_port_type_compact.wsdl">viewed
    online</a>.
</p>



_NIMBUS_CENTER2_COLUMN_END
_NIMBUS_FOOTER1
_NIMBUS_FOOTER2
_NIMBUS_FOOTER3