Abstract
Future nomadic communication will be enabled by huge TINA-like systems with millions of computational objects in it interacting with each other for the purposes of service provision and mobility management. In a very high degree, the nomadic customer will stay within a mobile cellular network and thus, will be connected to the fixed network part over a wireless link. However, up to now distributed platforms do not provide for any way at all to perform a binding to computational objects arranged in mobile devices. This paper presents an approach for realising CORBA-compliant wireless distributed systems that can be embedded in mobile cellular networks like today’s GSM or the future UMTS. The aspects of terminal mobility are hidden from the objects, and the existing infrastructure of these systems can be integrated without making any modifications.
Introduction: The Challenge of Nomadic Computing
In the late 1990s, the vision of "nomadic communication" predicted a world of millions of interacting software objects—based on frameworks like TINA and CORBA—providing services across global networks. However, these distributed platforms were fundamentally designed for fixed networks where hosts possess stable addresses and reliable connections.
Mobile devices introduced significant complexity: they move between cells, change network addresses, and suffer from intermittent wireless connectivity. At the time, no standard mechanism existed to "bind" or reliably invoke a software object residing on a mobile phone from a fixed network. This research proposes a solution to bridge this gap, treating mobile devices as first-class citizens in the distributed computing landscape without requiring an overhaul of existing cellular infrastructure.
The Bridging Architecture
The core of the proposal is a Bridging Architecture that mediates between wired domains (Internet/SS7) and the wireless domain. This architecture relies on several key components to maintain connectivity and protocol compatibility.
As shown in the protocol stack above, the system introduces the Wireless Inter-Operation Protocol (WIOP). WIOP is a specialized mapping of CORBA’s General Inter-Operation Protocol (GIOP) optimized for radio interface constraints. It utilizes a unique addressing format that includes a Temporary Mobile Terminal Identifier (TMTI) and radio channel information.
The Bridge itself resides in the access network and acts as a gateway. It employs Proxy Objects (POs) to transform requests from standard protocols, such as IIOP (Internet Inter-ORB Protocol), into WIOP. To manage these connections, the bridge maintains "Object Tables" and "Grouping Tables" to track reachable mobile objects.
Leveraging Cellular Infrastructure
A primary innovation of this approach is its integration with existing Intelligent Network (IN) and cellular components. Rather than reinventing location tracking, the bridge utilizes the established capabilities of GSM and UMTS networks.
When a request is made for a mobile object, the bridge interacts with the Service Data Function (SDF) and Paging Entities. It triggers a standard cellular "page" to locate the device's current cell within a specific Location Area. This allows for full location transparency: a fixed server can invoke a method on a mobile device without needing to know the device is mobile. The complexities of handovers and paging are entirely hidden behind the CORBA interface.
Key Insights and Efficiency
To address the high cost of radio bandwidth, the authors propose a Trading Proxy Service (TPS) located on the mobile device. Instead of each object on a device registering individually with the network, the TPS aggregates these "service offers" and transmits them to the bridge in a single burst. This grouping significantly reduces signaling overhead. Furthermore, the system supports both implicit registration (discovery via an outgoing call) and explicit registration (proactive advertisement to a global directory).
Conclusion and Future Outlook
The paper establishes a formal architectural framework and Interface Definition Language (IDL) specifications for WIOP. While the proposal is primarily theoretical and lacks quantitative performance data, it demonstrates a viable path for embedding CORBA-compliant systems into 3G (UMTS) networks.
Future work remains in the areas of empirical validation through mathematical modeling and simulations. Additionally, the scalability of maintaining proxy objects for millions of devices and the security implications of allowing external network objects to trigger paging on private terminals are critical areas for further investigation.
Summary updated