Source-linked AI summary
A gap analysis of Internet-of-Things platforms
Julien Mineraud, Oleksiy Mazhelis, Xiang Su, Sasu Tarkoma
TL;DR
IoT platforms must satisfy the needs of device providers, application developers, end-users, and other ecosystem participants, but existing solutions leave important capabilities incomplete. The paper evaluates a representative sample through a multi-axis gap analysis and a survey of Finnish IoT experts, then derives recommendations for filling the identified gaps. It emphasizes stronger marketplaces, developer support, local and flexible data handling, and improved access control and sharing mechanisms.
Problem
Existing IoT middleware solutions must meet diverse ecosystem participants’ expectations, but their readiness and usability across key platform capabilities require evaluation.
Method
The paper conducts a gap analysis of representative proprietary and open-source platforms and complements it with a survey of Finnish IoT program experts.
Results
The analysis and survey identify gaps in platform capabilities and support recommendations for extending contemporary IoT platforms.
Takeaways & Limitations
Recommended extensions include dedicated IoT marketplaces, SDKs and open APIs, local data analysis, flexible access control, and data processing and sharing mechanisms.
Takeaways & Limitations
The security and privacy analysis is limited to data-storage protection and data-access mechanisms available on IoT platforms.
Abstract
from arXiv · showhide
We are experiencing an abundance of Internet-of-Things (IoT) middleware solutions that provide connectivity for sensors and actuators to the Internet. To gain a widespread adoption, these middleware solutions, referred to as platforms, have to meet the expectations of different players in the IoT ecosystem, including device providers, application developers, and end-users, among others. In this article, we evaluate a representative sample of these platforms, both proprietary and open-source, on the basis of their ability to meet the expectations of different IoT users. The evaluation is thus more focused on how ready and usable these platforms are for IoT ecosystem players, rather than on the peculiarities of the underlying technological layers. The evaluation is carried out as a gap analysis of the current IoT landscape with respect to (i) the support for heterogeneous sensing and actuating technologies, (ii) the data ownership and its implications for security and privacy, (iii) data processing and data sharing capabilities, (iv) the support offered to application developers, (v) the completeness of an IoT ecosystem, and (vi) the availability of dedicated IoT marketplaces. The gap analysis aims to highlight the deficiencies of today's solutions to improve their integration to tomorrow's ecosystems. In order to strengthen the finding of our analysis, we conducted a survey among the partners of the Finnish IoT program, counting over 350 experts, to evaluate the most critical issues for the development of future IoT platforms. Based on the results of our analysis and our survey, we conclude this article with a list of recommendations for extending these IoT platforms in order to fill in the gaps.
1. Introduction
The article examines IoT platforms as middleware connecting smart objects to the Internet and evaluates how well they meet ecosystem participants’ needs. It combines a multi-stakeholder gap analysis with a Finnish IoT expert survey to identify deficiencies and recommendations.
- Study scope: The gap analysis evaluates today’s platforms against challenges arising from IoT development and ecosystem participants’ expectations.The article reviews platform capabilities and identifies gaps requiring attention in future platforms.
- Study scope: IoT platforms are studied as middleware and infrastructure enabling end-users to interact with smart objects.The study considers device vendors, application developers, platform and service providers, and end-users.
- Study scope: The Finnish IoT program survey strengthens the gap analysis by identifying critical gaps for future IoT platforms.The survey complements the platform-landscape evaluation with expert perspectives.
- Study outcome: The article concludes with recommendations intended to fill gaps identified through the analysis and survey.The paper’s later sections review platforms, analyze gaps, report survey results, and present recommendations.
- Current landscape: IoT platforms currently support interactions between devices and users, while interactions between platforms remain limited and costly.This frames interoperability between platforms as an unresolved aspect of end-to-end IoT interactions.
2. Review of today’s IoT platforms
The review surveys a representative, non-exhaustive sample of proprietary and open-source IoT platforms and compares characteristics relevant to users and application developers. It highlights widespread REST support but limited service discovery and continuing dependence on platform-specific connectivity choices.
- Platform sample: The review covers 39 proprietary and open-source IoT platforms as a representative, non-exhaustive sample.The platforms connect smart objects to the Internet and are listed with further details in Appendix A.
- Evaluation framework: Table 1 summarizes platform characteristics considered fundamental for meeting user and application-developer expectations.Its color coding provides quick visual information for platform selection.
- Device integration: Platforms requiring proprietary gateways depend on providers to adopt emerging technologies and support increasingly heterogeneous IoT devices.This dependence can limit reactivity to new protocols and device types.
- Platform types and architectures: Most platforms are cloud-provisioned as PaaS or SaaS, while architectures may be centralized or decentralized depending on deployment needs.The paper gives home environments and large sensor deployments as examples favoring different architecture choices.
- Openness: Open-source platforms are presented as promising because they may accelerate cross-domain integration and bottom-up technology adoption.The source frames these as expected or reported benefits rather than universal outcomes.
- Interfaces and discovery: Only a few platforms lack REST APIs, but only a few also integrate service-discovery mechanisms, even in simplified form.The review connects REST support with Web of Things tendencies and identifies discovery as comparatively limited.
Security and privacy of IoT platforms
Security and privacy are fundamental IoT-platform criteria, but the paper finds several unresolved challenges. Its analysis focuses specifically on data storage protection and data-access mechanisms, including access-control flexibility.
- Security requirements: Security and privacy are identified as fundamental criteria for IoT platforms.The paper situates these criteria within broader security and trust-management research.
- Unresolved challenges: Cloud-based IoT platforms are described as vulnerable to traditional attacks including denial of service, man-in-the-middle, eavesdropping, spoofing, and control attacks.The surveyed security literature also identifies device authentication, privacy, storage protection, trust management, governance, and fault tolerance as critical areas.
- Analysis scope: The paper limits its security and privacy analysis to data-storage protection and data-access mechanisms available on IoT platforms.Other security, privacy, and trust challenges are outside this analysis and referred to related surveys.
- Access control: Access-control flexibility ranges from basic private/public choices to fine-grained schemes allowing private, protected, public, or anonymous data.The paper regards the four-level scheme as the only surveyed option with sufficient flexibility to maximize data reusability by remote third-party services.
3. Gap analysis
The gap analysis finds substantial shortcomings across device interoperability, data ownership and privacy, data processing and sharing, developer support, and ecosystem integration. Current platforms provide partial mechanisms, but remain fragmented and limited in extensibility, control, analytics, discovery, and cross-platform development.
- Integration of sensing and actuating technologies: Device interoperability remains limited because platforms either support few technologies or require proprietary gateways, constraining heterogeneous-device integration.Some platforms use standard protocols, while others rely on proprietary gateways or require full-fledged HTTP endpoints.
- Data ownership: Cloud platforms rarely provide full data ownership: raw data is commonly uploaded, stored unencrypted, and protected with little disclosed security information.The analysis identifies this as Gap G2.1 and contrasts cloud storage with local-storage solutions.
- Data ownership: Access-control designs are often either overly strict or too permissive, leaving users without sufficient fine-grained control over data visibility.Permissions may be assigned through end-user interfaces, application implementations, or public feeds accessible without keys.
- Data processing and data sharing: IoT platforms have limited aggregation of data-processing techniques, while unreliable and incomplete streams make fault management and complete processing difficult.The paper points to data fusion and edge analytics as mechanisms for richer processing, improved energy efficiency, reduced privacy threats, and lower latency.
- Data processing and data sharing: Data discovery remains immature: only four reviewed platforms provide data-stream search, and cross-platform search is unavailable.Existing search is limited to platform-local streams, using mechanisms such as geolocation, tags, or data types.
- Support of application developers: Most platforms expose REST APIs, but nonuniform APIs and data models complicate data mashups across platforms.Only four studied platforms lacked a REST API, while many others also provide libraries, sometimes open source.
4. The perspective of the national Finnish IoT program
A survey of Finnish IoT program experts assessed business opportunities, risks, and important platform features. Respondents emphasized data reusability, ecosystem scale, and marketplace capabilities.
- Survey results: 40% rated maximizing data reusability very important, while 50% rated it quite important.
- Survey results: The most critical risks were lacking data suppliers and application providers and having too small a customer base.
- Survey results: A larger customer base attracts application developers and data suppliers, while those suppliers attract more customers.
- Survey results: Experts did not consider the marketplace’s possible negative impact on direct selling a critical risk, while landscape verticality was moderately critical.
- Priority features: Top features included publishing and describing applications or data, purchasing usage rights, and managing application sales, searches, registration, and subscriptions.
- Priority features: Eight of the ten most important features concerned IoT marketplaces and cross-platform interaction.
5. Recommendations for the development of IoT middleware
The recommendations combine short-term marketplace development with longer-term improvements to interoperability, data handling, developer support, ecosystem formation, and standardization. Dedicated marketplaces are presented as mechanisms for sharing data and applications across platforms.
- Recommendations: The recommendations cover heterogeneous-device protocols, local data processing, uniform data models and catalogs, edge analytics, streamlined APIs, cross-platform brokers, incentives, and dedicated IoT marketplaces.
- Short-term perspective: A basic IoT marketplace should initially serve as a repository for data streams and applications.
- Short-term perspective: Marketplace access to external streams would let users enrich local content and publish selected streams to third parties.
- Short-term perspective: Marketplaces would let application developers publish products and reach a wide range of customers, while supporting new innovations and business models.
- Short-term perspective: Marketplace functions would support discovering, marketing, selling, and purchasing rights to use data and applications.
- Recommendations: Fine-grained access control is recommended to restore users’ full data ownership, alongside SDKs that simplify application creation.
- Long-term perspective: Longer-term marketplaces could encourage uniform REST APIs, data models, and communication standards, while accounting could strengthen ecosystems and scale.
- Long-term perspective: The marketplace is positioned as a central connector enabling cross-platform data exchanges, business operations, and distribution of IoT applications.
6. Conclusions
The article evaluates a representative sample of proprietary and open-source IoT platforms through a multi-dimensional gap analysis and complements it with an expert survey. It concludes with short- and long-term recommendations addressing the identified gaps.
- Conclusions: The gap analysis assessed heterogeneous hardware integration, data management, application-developer support, ecosystem formation, and dedicated marketplaces.
- Conclusions: A Finnish IoT program expert survey evaluated business opportunities, risks, and important features associated with filling those gaps.
- Conclusions: The paper compiled short- and long-term recommendations intended to fill gaps in contemporary IoT platforms.
Appendix A. Reviewed IoT Platforms
The reviewed platforms span proprietary and open-source cloud, M2M, SaaS, and device-management solutions, offering varied connectivity, data, analytics, development, privacy, and ecosystem capabilities.
- Arkessa provides cloud-based M2M management, connects devices to many applications, retains data ownership for end-users, and offers a device-and-application ecosystem.Its MO-SAIC platform supports connections to many applications, while third-party privacy controls resemble those of Facebook or LinkedIn.
- EveryAware supports extendable data models, data-feed sharing and processing, REST access, and four levels of data visibility.The visibility levels are details, statistics, anonymous, and none.
- Other reviewed platforms illustrate diverse implementation and access models, including RFID handling, multi-source analytics, rule-based service composition, robot control, and constrained-device connectivity.Examples include Fosstrack, GroveStreams, IFTTT, MyRobots, and ARM mbed, with interfaces and data formats varying across platforms.
- H.A.T. makes individual data ownership central to a smart-home, multi-sided-market platform intended to create economic and business opportunities.The platform gives end-users control of their data to address privacy and related expectations.
- The Ericsson IoT-Framework focuses on analytics and data mashups, combining REST access, storage, OpenId access control, publish/subscribe, and data-stream querying.Its querying mechanisms cover local and external data sources for analytical tasks.