Source-linked AI summary
Survey of Communication Protocols for Internet-of-Things and Related Challenges of Fog and Cloud Computing Integration
Jasenka Dizdarevic, Francisco Carpio, Admela Jukan, Xavi Masip-Bruin
TL;DR
The rapid growth of IoT devices and data creates scalability and integration challenges for cloud, fog, and IoT architectures. This paper surveys application-layer protocols, compares interaction models and performance issues, and examines protocol deployment across system segments. It concludes that MQTT and RESTful HTTP are the most mature choices, while complete IoT-fog-cloud communication remains unlikely to rely on one protocol alone.
Problem
Growing IoT data volumes and fragmented communication protocols create an unresolved need for scalable, interoperable integration across IoT, fog, and cloud systems.
Method
The paper surveys application-layer protocols, compares request-reply and publish-subscribe models, reviews implementations across IoT, fog, and cloud segments, and examines performance and integration challenges.
Results
MQTT and RESTful HTTP emerge as the two most mature choices, while no single surveyed protocol is likely to cover communication from constrained devices to cloud servers.
Takeaways & Limitations
Integrated IoT-fog-cloud architectures should consider protocol combinations and interoperability because different system layers impose different communication requirements.
Takeaways & Limitations
Security mechanisms such as TLS and DTLS add overhead and can reduce protocols’ lightweight properties, while privacy remains under-addressed.
Abstract
from arXiv · showhide
The fast increment in the number of IoT (Internet of Things) devices is accelerating the research on new solutions to make cloud services scalable. In this context, the novel concept of fog computing as well as the combined fog-to-cloud computing paradigm is becoming essential to decentralize the cloud, while bringing the services closer to the end-system. This paper surveys on the application layer communication protocols to fulfil the IoT communication requirements, and their potential for implementation in fog- and cloud-based IoT systems. To this end, the paper first presents a comparative analysis of the main characteristics of IoT communication protocols, including request-reply and publish-subscribe protocols. After that, the paper surveys the protocols that are widely adopted and implemented in each segment of the system (IoT, fog, cloud), and thus opens up the discussion on their interoperability and wider system integration. Finally, the paper reviews the main performance issues, including latency, energy consumption and network throughput. The survey is expected to be useful to system architects and protocol designers when choosing the communication protocols in an integrated IoT-to-fog-to-cloud system architecture.
1 INTRODUCTION
IoT growth and expanding data volumes motivate architectures that can process and store data at scale, while protocol fragmentation complicates integrated IoT, fog, and cloud systems. The paper surveys application-layer protocols and their interoperability challenges across these segments.
- IoT device growth and the resulting data volumes motivate architectures that can process and store data at scale.
- Fog computing is presented as a way to scale and optimize IoT infrastructures beyond currently used cloud-based architectures.
- IoT, fog, and cloud deployments use fragmented protocols because their domains combine constrained devices, computing power, connectivity, security, and other requirements.
- The survey compares application-layer protocols and their request-reply and publish-subscribe characteristics, then examines implementations across IoT, fog, and cloud segments.
- The paper focuses on open issues involving seamless interoperability, coordination, and integration across combined IoT-fog-cloud architectures.
2 BACKGROUND
The background introduces fog-to-cloud architectures and application-layer communication through request-reply and publish-subscribe models. It contrasts their interaction patterns and summarizes protocol features relevant to IoT deployments.
- Fog-to-cloud systems add functional abstractions between IoT end devices and cloud services, including low-cost processing platforms such as Arduinos and Raspberry Pis.
- Request-reply model: Request-reply uses client/server message exchange in which a client requests information and a server processes and returns a response.
- Request-reply model: HTTP/2 adds asynchronous multiplexing over one TCP connection, while HTTP/1.1 pipelining requires responses in request order.
- Publish-subscribe model: Publish-subscribe uses publishers, subscribers, and brokers, allowing clients to subscribe to topics instead of repeatedly requesting information.
- Publish-subscribe model: Publish-subscribe supports decoupled many-to-many communication, allowing multiple publishers and subscribers to exchange updates without requiring simultaneous activity.
- Protocol comparison: MQTT, AMQP, and DDS implement publish-subscribe, whereas REST HTTP and CoAP implement request-reply; XMPP and HTTP/2 can support both models.
3 COMMUNICATION PROTOCOLS OVERVIEW
The surveyed application-layer protocols differ in interaction model, transport, quality-of-service options, security mechanisms, and suitability for constrained environments. HTTP and MQTT illustrate contrasting request/reply and publish-subscribe approaches, while CoAP, DDS, and XMPP address different integration and resource constraints.
- The survey compares protocols by standardization status, interaction model, QoS options, transport protocol, and security mechanisms.
- HTTP: HTTP uses RESTful request/response messaging over TCP, mapping POST, GET, PUT, and DELETE to resource operations.TLS enables HTTPS, but its handshake adds traffic during connection establishment.
- HTTP: HTTP’s TCP setup, long headers, power consumption, and request/reply model create challenges for constrained devices and push notifications.HTTP/2.0 addresses some concerns through compressed headers and concurrent exchanges over one connection.
- MQTT: MQTT uses lightweight publish-subscribe messaging with three QoS levels, trading resource requirements against delivery guarantees.QoS 0 is best effort, QoS 1 requires acknowledgment and retransmission, and QoS 2 delivers exactly once without duplication.
- DDS provides decentralized peer-to-peer publish-subscribe communication without a broker, while its security specification and adoption remain ongoing concerns.XMPP is disadvantaged in constrained or lossy networks by large XML messages, absent reliable QoS guarantees, persistent TCP, and inefficient binary encoding.
4 PERFORMANCE COMPARISON
The survey compares application-layer IoT protocols across latency, bandwidth, energy, security overhead, and adoption. Results vary by transport protocol, network conditions, payload size, and implementation context, so protocol selection remains scenario-dependent.
- Latency: Latency comparisons favor different protocols under different conditions: MQTT often outperforms HTTP, while CoAP can outperform MQTT depending on packet loss, QoS, and network conditions.MQTT had lower latency than HTTP in fog-to-cloud and home-automation scenarios; CoAP showed lower RTT or latency in several other conditions.
- Latency: For broker-based protocols, MQTT and AMQP have nearly equal latency with small payloads, but MQTT has lower latency with huge payloads.
- Latency: HTTP/2 performs better than CoAP under high congestion, whereas CoAP performs better under lower congestion.
- Latency and security: TCP-based MQTT, AMQP, HTTP, and XMPP generally exhibit higher latency than CoAP and UDP-based DDS, while security layers add 6.5% TLS and 11% DTLS overhead.The survey attributes this difference to the underlying transport protocol and identifies security overhead as especially important for IoT-to-fog communication.
- Bandwidth consumption and throughput: CoAP generally reduces bandwidth use and improves efficiency relative to MQTT, especially for small messages and dynamic environments with many simultaneous devices.Protocol efficiency is defined as useful information bytes divided by total bytes exchanged at the application and transport layers.
- Energy consumption: Energy results favor MQTT over HTTP, CoAP over MQTT in constrained gateways, and MQTT over AMQP in mobile or unstable wireless networks.CoAP is also reported to have efficient power-management capabilities, while AMQP provides more security-related aspects in one comparison.
- Developer’s choice: MQTT and HTTP are the most adopted protocols because developers view them as more mature and stable IoT standards.
5 CONNECTING THE IOT, FOG AND THE CLOUD
The section examines how application protocols can be positioned across IoT, fog, and cloud layers, including single-protocol and combined-protocol architectures. It emphasizes lightweight protocols for constrained IoT devices, interoperability challenges, and MQTT-based integration examples.
- Architecture and integration: No communication protocol was originally designed for combined IoT-fog-cloud systems, leaving protocol placement and interoperability as open research areas.All surveyed protocols operate at the application layer and could theoretically serve different system segments.
- Single- and multiple-protocol solutions: Single-protocol solutions are easier to implement, but individual protocols optimally satisfy different requirements across IoT, fog, and cloud layers.MQTT and REST HTTP are identified as widely used candidates for single-protocol solutions, while combined architectures can position protocols by layer.
- MQTT example: An MQTT smart-farm example maps Raspberry Pi sensor clients to a fog broker and can bridge that broker to a cloud broker.A mobile application may subscribe to both brokers and fall back from the cloud broker to the fog broker if the cloud connection fails.
- Protocol placement: IoT-to-fog communication commonly uses MQTT, CoAP, or XMPP, whereas fog-to-cloud communication commonly uses AMQP and REST HTTP.RESTful HTTP and DDS are also described as potentially usable across all layers.
- Protocol placement: Lightweight operation on constrained devices is the key suitability criterion for IoT-to-fog communication, while fog-to-cloud links need not satisfy that constraint.The proposed direction is to combine lightweight protocols below with protocols unrestricted by constrained-device requirements above.
- Request-reply integration: HTTP and CoAP can be placed in different layers and connected through proxies that translate requests between HTTP clients and CoAP servers, or vice versa.REST HTTP is presented as reasonable for more powerful fog-cloud nodes, while CoAP suits resource-constrained IoT devices.
- Open challenges: Security mechanisms such as TLS and DTLS can address some security concerns but may reduce the lightweight properties of communication protocols.Research on adapting these mechanisms remains ongoing, while privacy remains an unresolved concern.
6 CONCLUSIONS AND OUTLOOK
The conclusion identifies MQTT and RESTful HTTP as the most mature protocol choices, while arguing that complete IoT-to-fog-to-cloud communication generally requires protocol selection by system segment. It also highlights unresolved standardization, HTTP sufficiency, security, and privacy issues.
- Conclusions: The survey finds that no single surveyed protocol is likely to cover communication from resource-constrained IoT devices through fog systems to cloud servers.Different requirements across IoT, fog, and cloud motivate considering multiple protocols.
- Conclusions: MQTT and RESTful HTTP are identified as the two most mature and developer-favored choices, supported by stable implementations and extensive documentation.MQTT is associated with constrained IoT devices, while RESTful HTTP is a straightforward choice where battery consumption and constrained communication are less important.
- Outlook: Defining standards to unify varying architectures and interfaces remains a major challenge for combined IoT, fog, and cloud management.The integration of fog computing with existing IoT-cloud proposals is described as a novel proposition with multiple architectural possibilities.
- Outlook: RESTful HTTP is not sufficient for combined IoT, fog, and cloud solutions according to performance studies, leaving future protocol choice unresolved.The paper leaves open whether future work will adopt other messaging protocols or improve HTTP.
- Outlook: Security and privacy require further analysis, particularly regarding the overhead introduced by security mechanisms.TLS and DTLS can affect lightweight properties when combined with surveyed protocols.
GLOSSARY
The glossary defines abbreviations for communication, security, networking, and Internet-of-Things concepts used throughout the paper.
- Glossary: The glossary expands protocol and security abbreviations including AMQP, CoAP, DDS, DTLS, HTTP, MQTT, REST, TLS, and QUIC.It also includes IoT, IETF, JSON, SASL, GDS, and SPDY.