Decentralized Stream Processing Over Web-Enabled Devices

Question

1. What have the authors contributed in "Decentralized stream processing over web-enabled devices" ?

2. What are the future works in "Decentralized stream processing over web-enabled devices" ?

3. What is the function that spreads the temperature across the available Peers?

4. How many parallel operators can be added to the pipeline?

Accepted Answer

In this paper the authors present the Web Liquid Streams framework for building and executing stream processing topologies capable of gathering data from Web-enabled sensors and process it through JavaScript operators scattered across a peer-to-peer Cloud of computing peers.. The authors present the platform and its programming interface, showing a first evaluation of the system.

Accepted Answer

The authors plan to further studying the robustness, scalability and performance of WLS by porting some stream processing benchmarks to JavaScript [ 3 ], and by using WLS to build more complex Topologies in real-world use case scenarios.

Accepted Answer

The Web Liquid Streams framework runtime is able to spread the Operators across the available Peers following the instructions of the JSON topology description file.

Accepted Answer

The authors can see that by increasing the number of Peers the system can increase the number of parallel filter operators and further reduce the message processing delay.

Accepted Answer

The authors also demonstrated that WLS can use the most appropriate communication protocol to stream data between peers and compared the corresponding overhead.

Accepted Answer

The Web Liquid Streams framework helps developers create stream processing topologies and run them across a peer-to-peer Cloud of connected devices, where they share and make use of shared resources.

Accepted Answer

The first Producer Operator generates a finite list of numbers which are sent one at a time downstream with a fixed data rate of 10 messages per second.

Accepted Answer

To run an Operator, a Web browser has to connect to a specific path (that is, performs a GET request) in a Web server which is running an instance of WLS.

Accepted Answer

Given an adequate amount of resources, deployment can happen on browsers only without performance loss,8 16 24 32 40 48 560.80.91Number of Parallel Filter OperatorsR ela tiv eT h rou g h p u tNode.js Web Browser Both(a) Relative Throughput obtained with different deployment configurations having a larger pool of peers8 16 24 32 40 480204060Number of Parallel Filter Operators

Accepted Answer

If a Topology is making use of the Peer about to leave, the operators running on it are automatically migrated on another available Peer.

Accepted Answer

The authors have chosen to analyse the simplest Topology with a toy example in order to reduce complexity and obtain repeatable results about the deployment differences of a single Operator.

Accepted Answer

The authors plan to further studying the robustness, scalability and performance of WLS by porting some stream processing benchmarks to JavaScript [3], and by using WLS to build more complex Topologies in real-world use case scenarios.

Accepted Answer

There is no constraint on the data structure of the messages exchanged along a Topology, as long as they can be serialized into JSON strings.

Accepted Answer

This evaluation shows that even deploying a Topology on a very powerful machine, by using enough weak machines the authors can achieve a better performance.

Accepted Answer

In the first example, by increasing the throughput at the producer Operator, the authors increase the computational effort on the Topology and show how the system performs by parallelising on a set of finite resources.

Accepted Answer

The results (Figure 3) show that for a small webcam feed (320x240) the authors can reach a throughput of around 35 to 40 frames per seconds.

Accepted Answer

Topologies may also be modified at runtime by, for example, adding/removing one or more Operators, resulting in a structural change of semantics of the Topology itself, which will thus affect its end result.

Decentralized Stream Processing Over Web-Enabled Devices

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. What have the authors contributed in "Decentralized stream processing over web-enabled devices" ?

2. What are the future works in "Decentralized stream processing over web-enabled devices" ?

3. What is the function that spreads the temperature across the available Peers?

4. How many parallel operators can be added to the pipeline?

5. How did the authors demonstrate the performance of WLS?

6. What is the main abstract stream processing framework?

7. How many messages are sent to the producer operator?

8. What is the architecture of the client-side WLS?

9. How many parallel filters can be deployed on a browser?

10. What happens if a Topology is making use of the Peer about to leave?

11. Why did the authors choose to use a toy example?

12. How do the authors study the performance of WLS?

13. What is the constraint on the data structure of the messages?

14. How does the evaluation show that a topology can be deployed on a very powerful machine?

15. How does the system perform when the data throughput changes?

16. How many frames per second can be achieved?

17. What can be done to modify a topology at runtime?

Figures

Citations

Service-Oriented and Cloud Computing

LiquidAI: Towards an Isomorphic AI/ML System Architecture for the Cloud-Edge Continuum

Service-Oriented and Cloud Computing: 9th IFIP WG 6.12 European Conference, ESOCC 2022, Wittenberg, Germany, March 22–24, 2022, Proceedings

Mashup Development with Web Liquid Streams

Liquid stream processing on the web: a JavaScript framework

References

Processing flows of information: From data stream to complex event processing

Gossip-based aggregation in large dynamic networks

MillWheel: fault-tolerant stream processing at internet scale

Discretized streams: an efficient and fault-tolerant model for stream processing on large clusters

A Web of Things Application Architecture: Integrating the Real-World into the Web

Related Papers (5)

Integrating multimedia streams into a distributed computing system

Design of composable proxy filters for heterogeneous mobile computing

Active Streams - an approach to adaptive distributed systems

Distributed file sharing using web services

Management of real-time streaming data Grid services: Research Articles