The IBM wireless sensor networking testbed

Furrer, Simeon; Schott, W.; Truong, Hong Linh; Weiss, Beat

doi:10.1109/tridnt.2006.1649126

Cited by 14 publications

(5 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…GStreamer [7] and its off-the-shelf plugins are still the basis of this work. For capability-based publish/subscribe stream connections, we adopt MQTT [4] via "paho.mqtt.c" library [3]. We have started initial prototypes with TCP and RTSP plugins.…”

Section: Designmentioning

confidence: 99%

“…Then, we have applied Ze-roMQ [9], which, fortunately, has an off-the-shelf and open sourced GStreamer plugin for second prototypes. Then, we have switched to MQTT [4] for the last prototypes. We choose MQTT because, ultimately, we want to contribute to IoT standards including Matter [1] and SmartThings [20], which include MQTT.…”

Section: Designmentioning

confidence: 99%

“…We may also have a pipeline that needs to connect to any compatible pipeline available or does not want to specify network addresses. Thus, after the initial prototype with ZeroMQ [9], we have implemented MQTT [4] stream pub/sub plugins compatible with general GStreamer data streams. With MQTT pub/sub plugins, we have implemented prototype With this prototype instance, the client's initial requirements (AI application developers) have been satisfied.…”

Section: Designmentioning

confidence: 99%

“…AI application and service developers have requested a similar capability to publish services and subscribe to the services. After iterations of different implementations from ROS, ZeroMQ [9], and MQTT [4], we have decided to use MQTT as the basis. ROS is not chosen because it is over-complicated for the need, requires recompiling software for updated schema, and excessively requires software packages for its dependencies.…”

Section: Protocolsmentioning

confidence: 99%

“…To satisfy the new requirements, we first extend the AI stream data type ("other/tensors" MIME) to support tensors with flexible dimensions and schemaless data streams. Then, in addition to off-the-shelf networking plugins including TCP, UDP, and RTSP, we propose to use MQTT [4] for service discovery and connections, which is applied to both publish/subscribe and workload offloading, along with synchronization methods. We also provide a lightweight library with minimal dependencies, "NNStreamer-Edge", which allows implementing non-NNStreamer software packages compatible with the proposed among-device AI pipeline connections.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Toward Among-Device AI from On-Device AI with Stream Pipelines

Ham¹,

Woo²,

Jung³

et al. 2022

Preprint

View full text Add to dashboard Cite

Modern consumer electronic devices often provide intelligence services with deep neural networks. We have started migrating the computing locations of intelligence services from cloud servers (traditional AI systems) to the corresponding devices (on-device AI systems). On-device AI systems generally have the advantages of preserving privacy, removing network latency, and saving cloud costs. With the emergent of on-device AI systems having relatively low computing power, the inconsistent and varying hardware resources and capabilities pose difficulties. Authors' affiliation has started applying a stream pipeline framework, NNStreamer, for on-device AI systems, saving developmental costs and hardware resources and improving performance. We want to expand the types of devices and applications with on-device AI services products of both the affiliation and second/third parties. We also want to make each AI service atomic, re-deployable, and shared among connected devices of arbitrary vendors; we now have yet another requirement introduced as it always has been. The new requirement of "amongdevice AI" includes connectivity between AI pipelines so that they may share computing resources and hardware capabilities across a wide range of devices regardless of vendors and manufacturers. We propose extensions of the stream pipeline framework, NNStreamer, for on-device AI so that NNStreamer may provide among-device AI capability. This work is a Linux Foundation (LF AI & Data) open source project accepting contributions from the general public.

show abstract

Section: Designmentioning

confidence: 99%

Section: Designmentioning

confidence: 99%

Section: Designmentioning

confidence: 99%

Section: Protocolsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations