Masterarbeit
Dynamically Partitioning WebAssembly-based Client-Server Applications
Completion
2025/01
Research Area
Students
Advisers
Dr.-Ing. Sebastian Heil
Description
In distributed web applications using a client-server architecture the application logic is partitioned between client- and server-side, with some parts running on the client and others on the server. This partition is traditionally decided during design time by the developers of the application. Given the heterogeneous nature of devices using the web, especially considering the vastly differing capabilities of mobile devices, this one-size-fits-all-approach is not ideal, since it does take neither the devices’ capabilities nor the dynamically changing situation, such as system load, network state, or remaining battery, into account. This can lead to degraded performance on weaker client devices, or an increased server load despite an under-utilized client, both of which are not desirable for a service provider.
The idea of the following thesis is the creation of a dynamic decision-making framework, which allocates parts of the application logic between multiple clients and a server based on dynamically measured system and network information. To achieve this, the required information needs to be measured on both client- and server-side during runtime, and continuous adjustments made to the client-server partition of application logic. Given that clients are not directly under control of the service provider, the impact of clients with reduced functionality or which intentionally disrupt the measuring process, should be considered.
The objective of this thesis is the creation of a solution or the combination of existing approaches to solve the above problem of dynamic partitioning of client/server application logic in web applications through a measurement- and policy-based approach and infrastructure as described above. This comprises the analysis of the state of the art of code mobility, load balancing, and application partitioning, as well as the demonstration of the solution by prototypical implementation and a suitable evaluation based on experimentation simulating different load scenarios and quality/performance measurements.
