Optimizing a CORBA Internet inter-ORB protocol (IIOP) engine forminimal footprint embedded multimedia systems

摘要

To support the quality-of-service (QoS) requirements of embeddednmultimedia applications off-the-shelf middleware like common objectnrequest broker architecture (CORBA) must be flexible, efficient, andnpredictable. Moreover, stringent memory constraints imposed by embeddednsystem hardware necessitates a minimal footprint for middleware thatnsupports multimedia applications. This paper provides threencontributions toward developing efficient object request broker's (ORBs)nmiddleware to support embedded multimedia applications. First, wendescribe optimization principle patterns used to develop a time andnspace-efficient CORBA inter-ORB protocol (IIOP) interpreter for thenadaptive communication environment (ACE)-ORB (TAO), which is ournhigh-performance, real-time ORB. Second, we describe the optimizationsnapplied to TAO's interface definition language (IDL) compiler tongenerate efficient and small stubs/skeletons used in TAO's IIOP protocolnengine. Third, we empirically compare the performance and memorynfootprint of interpretive (de)marshaling versus compiled (de)marshalingnfor a wide range of IDL data types. Applying our optimization principlenpatterns to TAO's IIOP protocol engine improved its interpretiven(de)marshaling performance to the point where it is now comparable tonthe performance of compiled (de)marshaling. Moreover, our IDL compilernoptimizations generate interpreted stubs/skeletons whose footprint isnsubstantially smaller than compiled stubs/skeletons. Our resultsnillustrate that careful application of optimization principle patternsncan yield both time and space-efficient standards-based middleware