Integrating Messaging Middleware and Information Flow Control (original) (raw)
Related papers
2015
The SMS gateway EMG uses separate threads for each connection. This requires a large number of locks to avoid data corruption, and causes problems when scaling to thousands of connections. The CPU ...
A Low Cost and Resilient Message Queuing Middleware
2008
Message Queuing Middlewares (MQMs) are gaining more and more attention in large enterprises for building highly available asynchronous messaging systems and for integrating heterogeneous applications. However, currently available MQMs consider underlying networks as static. Therefore, in case of node failures or a disaster, either they have to suffer long term service loss or they need to install a lot of extra resources to ensure that no such failures cause any service loss. They also require large administrative overhead as the network is managed manually. Besides, as store-and-forward method is used, reliable delivery of messages suffers much network delay and generates large amount of traffics. Current MQMs are not suitable especially if the network contains a large number of nodes. In our previous work, we proposed a general purpose middleware called Soft System Bus (SSB) to solve the continuous availability problem. In this paper, we redesign SSB based on Pastry so that it solves the problems of large-scale MQMs. This middleware provides asynchronous, reliable and in-order delivery service while ensuring no long term service loss in case of failures or disasters. Such services can be provided with minimum deployment cost. Our simulation based evaluation shows that we can provide such services in a network of large number of nodes while generating less traffic and requiring minimum administrative overhead.
Extending Message-Oriented Middleware
Abstract: Different types of middleware exist to facilitate the integration of software running on heterogeneous computing platforms. Message Oriented Middleware (MOM), for instance, enables the interaction between heterogeneous applications by exchanging packets of structured data (messages) through communication channels. The core responsibility of a MOM is asynchronous delivery of messages from senders to receivers, as well as management of the corresponding message queues.
Adding Resilience to Message Oriented Middleware
2014
This paper describes an approach to enhance the resilience of message oriented middleware to faults and attacks. The system is designed to work with different message oriented middleware as long as the messaging system can supply the information and control API required. The structure of the system and the approach to handling link failures, broker failures and anomalous behaviour through the planning of mirroring and relaying mechanisms to suit the tolerable risk and the deployment context is the focus of the paper. Validation is through industrial case studies and end user assessment.
Behavior and Performance of Message-Oriented Middleware Systems
2002
The middleware technology used as the foundation of Internet-enabled enterprise systems is becoming increasingly complex. In addition, the various technologies offer a number of standard architectures that can be used by designers as templates to build applications. However, there is little concrete understanding in the software industry on the strengths and weaknesses of competing technologies, and the different trade-offs that various component architectures impose. The SACT Group at CSIRO has qualitatively and quantitatively evaluated a number of commercially available Message-Oriented Middleware (MOM) systems. This paper focuses on the results obtained from the performance evaluation of the IBM's MQSeries V5.2. It presents an overview of the technology, and discusses the metric used in this study for performance measurement The test results related to the sustainable performance of the system using various test configurations are described and their implications discussed.