Automatic Deployment of Transcoding Components for Ubiquitous, Network-Aware Access to Internet Services

Abstract

Advances in wireless communication together with the growing number of mobile end devices hold the potential of ubiquitous access to sophisticated internet services; however, such access must cope with an inherent mismatch between the low-bandwidth, limited-resource characteristics of mobile devices and the high-bandwidth expectations of many content-rich services. One promising way of bridging this gap is by deploying application-specific components on the path between the device and service, which perform operations such as protocol conversion and content transcoding. Although several researchers have proposed infrastructures allowing such deployment, most rely on static, hand-tuned deployment strategies restricting their applicability in dynamic situations.In this paper, we present an automatic approach for the dynamic deployment of such transcoding components, which can additionally be dynamically reconfigured as required. Our approach relies on three components: (a) a high-level integrated type-based specification of components and network resources, essential for "late binding" components to paths; (b) an automatic path creation strategy that selects and maps components so as to optimize a global metric; and (c) system support for low-overhead path reconfiguration, consisting of both restrictions on component interfaces and protocols satisfying application semantic continuity requirements. We comprehensively evaluate the effectiveness of our approach over a range of network and end-device characteristics using both a web-access scenario where client performance is for reduced access time, and a streaming scenario where client preference is for increased throughput. Our results verify that (1) automatic path creation and reconfiguration is achievable and does in fact yield substantial performance benefits; and (2) that despite their flexibility, both path creation and reconfiguration can be supported with low run-time overhead.

Java active extensions: Scalable middleware for performance-isolated remote execution

“…The extension uses the message-passing mechanism provided by the extension server to send to the endpoint messages containing the IP address of the machine on which it is executing and the ports the extension was able to open (Listing 2, lines [13][14][15]. This messagepassing mechanism is provided by the extension server for exactly this purpose of bootstrapping application-specific communication.…”

“…CANS provides an infrastructure for data adaptation that supports the insertion of application-specific components along the data path [15]. The execution environment they propose for intermediate nodes is tailored to their compos-able, component-based model.…”

Fractal: A Mobile Code Based Framework for Dynamic Application Protocol Adaptation in Pervasive Computing

“…Conductor [56] and CANS [14] provide an application transparent adaptation framework that permits the introduction of arbitrary adaptors in the data path between applications and end services. While these approaches provide an extremely general adaptation mechanism, significant change to existing infrastructure is required for their deployment.…”

“…Adaptation has been considered as a general approach to address the mismatch problem between clients and servers [14,24,36,51]. From the perspective of adaptation locations, some of them propose the in-network adaptation, such as CANS [14], Active Names [51], Odyssey [36], and Rover [24], which focus on how to do the adaptation step by step across an overlay path. Although the functionalities are well designed, they have not considered the deployment of chosen components (drivers in CANS [14]) across multiple nodes in the path.…”

SEE: a service execution environment for edge services

“…Although several infrastructures have been proposed [5,6,7,15] to do this, there has not been a widespread use of them due to the concerns about their deployability, performance, and scalability. Central to each of these concerns are the questions such as where to deploy these services, how to integrate them within the existing data flows, and who would select the services.…”