The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada Designing the Fog: Towards an Intranet of Things Mathias Funk Abstract Industrial Design Department The Fog of Things, ubiquitous computing in local Eindhoven University of Technology contexts, is a reality now. The Internet of Things has Eindhoven, the Netherlands arrived, although users use and perceive it rather as m.funk@tue.nl Internet of Thing [sic!]. Data and information flows vertically, not horizontally through the connected Everyday and promises of convenience and quality of life are at the mercy of the viability of business models and the benevolence of multi-national commercial entities. This position paper poses that things and connectedness can also be re-thought; products and services can be designed differently: bottom-up and with stronger ideals in place. Systems of connected things can be understood as horizontal autonomous networks of nodes, Intranets, that do not or only seldom connect to external entities for the exchange of data. This paper explains how to conceptualize these systems, proposes the use of system properties to Copyright © 2018 for this paper held by its author(s). Copying permitted address new design challenges, and concludes with an for private and academic purposes. outlook on future work. Author Keywords Smart things; Internet of Things; Systems Design; Interaction Design. ACM Classification Keywords H.5.m. Information interfaces and presentation. 31 The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada Introduction older devices, or even new investments. Finally, an Increasingly computing extends to the fringes of the installation of connected devices grows over time [5], Internet, towards local contexts that are in close due to changing needs, maintenance, or replacements proximity to users and that allow for direct interaction [6]. Sometimes a new device is added because it and consumption of ubiquitous services offered by the promised new functions that might or might not be computing infrastructure. Devices emerged that complementary to what already exists in the context. package new functionality based on connectedness, on This growth adds to the complexity, compatibility information streams and frequent data exchange [1]: issues, functionality overlaps, data that cannot be tracking, monitoring and measuring, notifying or easily exported and imported again, thereby alerting, adjusting and connecting, to name a few. This complicating the operation of a smart home. Especially has not only implications for new devices and services this last point shows that consumer IoT can become in that enter our personal spaces [2], there are principle more complex than industrial IoT installations, implications for how we see ourselves reflected in such which are carefully planned, designed and built to data, often freely accessible and exploitable, now or in clearly specified needs. The emergence of connected the future [3]. devices has clearly raised concerns about privacy and (data) ownership with users [7, 8]. Furthermore, Consumer IoT products may disguise as traditional devices are often implemented without proper security interactive products, but they are inherently more than mechanism when, for instance, accessing a local that: they are connected. And this has serious wireless access point or communicating data to a cloud implications [4]. They require not only more expertise appliance. When operating in a technologically and skills about networked technology to setup and uncertain or diverse environment such as homes, install, configuration of connected products is a offices, and public buildings, commercial IoT devices disguised maintenance process (similar to what is and systems embody a trade-off between ease of known from industrial or professional contexts). For deployment (not even ease of use) and data security. example, in the context of a smart home with tens of Usually, compromises on deployment would lead to connected devices, changing the Wifi password higher costs, so data security is likely to be neglected. becomes a task that might take days instead of Finally, IoT devices might easily live longer than their minutes, requiring the users to track back obscure manufacturers. Consumers expect a 5 to 10-year life, configuration methods and parameters – often per while the manufacturer might be a small company, device. There is the question of compatibility that is seed funding initiative or even side-project, and might beyond sockets, cables and plugs; compatibility no not exist after one or two years. By now, we have seen longer guaranteed by international industry-wide large corporations shedding IoT products from their standards, compatibility becomes a challenge at the portfolios the moment they become misaligned with higher level of APIs and connection protocols. These are grand strategy [9]. From that point onwards, IoT certainly out of control of users, and often change over devices that rely on external services might become time, which requires maintenance, workarounds for useless, or at least will not get updates anymore that 32 The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada would ensure compatibility and appropriate functioning From Internet of Thing to Intranet of Things and prevent hacking or misuse by adversaries [10]. It is important to understand that IoT technology for consumers in personal spaces is currently still in its The abovementioned challenges are well-known for infancy and will exhibit flaws that become apparent year and attractive for design. However, in the context only over time when technology, business models and of design, systems are often misunderstood and services evolve, when the true needs and expectations misrepresented [11, 12]. Furthermore, end-to-end of consumers become clear. Such a context with the systems design is often not practiced. Instead, the role premise of connectedness and ubiquitous computing of design is limited to interface-related usability and embedded into the environment is ideal for design: yet, user experience problems, and pure form-giving. When instead of technological frameworks for systems, we investigating available technology in the consumer need design frameworks for systems. In this paper, we space, a clear anti-pattern can be observed: The propose the design of connected things that will work Internet of Thing. These are small-scale systems, often independently of external services, that will continue to thing-app tandems, which occupy the personal as a serve without an Internet connection, and that will not singular product extending an invisible, intangible link share collected data with external parties. to the cloud. Examples are everywhere to be found, Internet of Thing devices that only work when they are Design space connected to servers in the cloud, devices that will not If we want to predict a second or third wave of work without centrally managed licenses and keys, and connected products that better behave, how can this control structures that are highly optimized for design space be characterized? From a technological performance, cost and control, but not for robustness angle, what we describe here is an intranet. Devices in and participative data ownership. Even just a tight an Intranet of Things form a constellation and might coupling between a physical “smart object” and a use the Internet as an optional in-bound source of data, proximal smart phone app extends the object’s but should not rely on it. Instead of extending their capabilities, but with the effect of centralizing control scope towards external services, they extend further towards the smart phone and encasing “smart” horizontally: from room to floor to household, from functionality in a silo of specifically this object-app street to neighborhood to community different scopes tandem. For systems design, these are clearly anti- are imaginable. Apart from this spatial characterization, patterns of well-designed systems. the scope encompasses an oversee-able collective of actors, people and things. In the remainder of this position paper, we will explain the main challenges for systems design in the context If we think of designing connected things that form a of IoT, and focus on emergent properties in systems of local constellation, the complexity of conceptualizing, things. The position paper continues with a discussion designing, deploying and using becomes larger – and concludes with a summary and a section on future especially if not all devices “play together” nicely. If we work. want to design for a future when devices will primarily 33 The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada exist in (local) constellations, we need to change (1) Things how designers look at products and users or We assume that all things in a constellation are, to stakeholders in the Intranet of Things, and (2) the way some degree, capable of computing data and using the designers understand and leverage technology fully. result in action and communication. Some things might focus on the sensing and transmission of sensor data. For (1), future connected products in a constellation of Other components fulfill different roles, which require connected things, are intended to be more than just the them to process data, filter events, store information, sum of each individual thing. This is currently not the manipulate weights of a learning algorithm and trigger case. The reason for connectivity is that products are actions. While these capabilities require different extensions of a cloud-based business processes, not configurations in computing power and storage, energy autonomous actors on their users’ behalf. Our design and connectivity might be impacted as well. In general, vision is that products are created as independent in the scope of this position paper, we assume that actors that can act in the local context based on shared sufficient computation is available at all layers of a information and semantics that are exchanged at a system. When a design moves towards production, local level. They act with autonomy that is a well- such requirements can be trimmed to better fit designed trait, not a fallback mechanism, nor after- production or business constraints. thought. The consequence is that design needs to extend beyond the scope of a single product and take This structure is, however, not enough; like cells of an ecologies [13] into account. organism, without exchange of fluids, decay sets in. Data and information flows are necessary to turn a For (2) and the foreseeable future, designers need to static set of components into a live system. The bridge an important gap: on the one side, there is IoT components are the essence from which systems technology built and deployed in the context that emerge. They determine the nature of the system, it’s follows the prevalent cloud connectivity paradigm, and “product-ness”, and specific tool or service qualities. on the other side, a new focus on locality and data The question what we are designing needs to start with ownership needs to be nurtured through new components, rather than the big picture, and then build technology which promotes different use cases and connections, layers, interconnections, and cross- thus different designs. To bridge this gap, a better, sections towards the emergence of system properties. more practical definition of systems of connected things is needed for approach better technological framings. Systems of Connected Things As described above, a challenge of systems design is In the following, we will first elaborate the “Intranet of that systems are often mischaracterized and thus Things” as a design vision, and then take a more oversimplified already in early phases of the design technological perspective towards thing ecologies as process. Simplifications target common convergence designed and evolved constellations of things. points: single products that are functionally overloaded, imbalanced product-app tandems and cloud services 34 The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada with product as crippled physical manifestations. To to realize these traits. The properties in Table 1 are counter this, product design needs to embrace the connected and non-contradictory, i.e., they are concept of “emergent properties”, i.e., desirable formulated from a user point of view and some of them properties of a system that a designer might find can only be realized in connection to other properties applicable in a design project. In the following, we being apparent in the design. The properties are describe a non-exhaustive selection of such emergent divided into three categories, structure, behavior, and properties (see Table 1). These properties demonstrate interfaces. For each category, two or three emergent balance between describing desirable traits of a system properties are presented and briefly discussed in the and relating to possible approaches in systems design context of the category. Property Application / Example Structure Growth Changing structure (extension, pruning, re-organization) Shared functionality Functional grouping Behavior Emergence through homogeneity Complex behavior: synchronization, self-organization Emergence through heterogeneity Social behavior (amongst things) Cooperation and Negotiation Dynamic roles, task switching Autonomy Self-contained functionality, independent operation Distributed cognition Shared sense-making, exploration of context, new situations Interfaces Adaptation through local learning Contextualization as adaptation to context Personalization as adaptation to user(s) (Inter-)locality and Representation Sharing information (amongst things) Sharing information (towards user(s)) Intentionality Social behavior (towards user(s)) Table 1: Overview of System Properties categorized into structure, behavior, and interfaces. Structure groups facilitate other emergent behavior such as When a system as a constellation of things grows, this shared functionality, emergence and redundancy. When means that its structure–often called (physical) graph– things connect, they can also establish hierarchies and changes: either the set of nodes is changed by adding, functional groups to address external requests, or to removing or replacing a node, or the linking between balance a heavy load. nodes changes. Such mechanisms of growth might result from a user purchasing a new device, a faulty Behavior device dropping out, or a firmware update. Growth in We know emergence from highly homogeneous how things connect is more intricate: things can systems in nature, e.g., fireflies synchronize their connect based on their spatial location and proximity, blinking and swarms of birds form and scare away they can form functional groups and through such predators that are larger than individual birds in the 35 The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada swarm. Such emergence can be leveraged in design to functionality and structural composition. Data can be synchronize concurrent processes in distributed raw measurements, control data or commands, or high- products. At the same time, product can engage in level information that influences the experience and social interaction based on their heterogeneity. Patterns decision making of users. As a system property, data of cooperation and negotiation can help further self- refers to sharing of data between things, or the sharing organization, decision making and recovery from fast of information between the system and its users. While growth. These behavioral patterns can be sharing data with components of a system obviously complemented by highly autonomous behavior that requires formal interfaces and clear protocols, sharing helps a thing maintain its function during a period of data with human users through data representations limited connectivity to its peers. This means that the such as visualizations, decorations, physicalizations or functionality of the system is not only distributed across even sonifications, requires a similarly careful several components, it means that this functionality is approach. When looking at interfaces and exchanges of performed regardless of central coordination in a information, the focus shifts towards the data that is redundant, parallel and autonomous way. Such a gathered, processed and stored in the system, where design can mitigate the failure of one component which data flows and whether data flows can be limited to a can be balanced out by another; components can local, yet systemic scope without leaking out “team up” to autonomously recognize failure of one unintentionally. This property, as a consequence, team member and act upon it, and functionality can be implies that designers need to consider how locality can provided in different locations simultaneously without be achieved without a central and global controller coordination. Finally, some systems might exhibit forms instance. To counter the need to involve global or of distributed cognition when utilizing different central control structured, designers can take a hybrid complementary capabilities of things to make sense of approach and consider inter-locality, i.e., a group of a new situation or context. “local” components that cooperatively share data or, more appropriately, models derived from data, that Interfaces abstract from local details whenever benefits of sharing When things interface with each other, or the context, such information between systems or subsystems they can be designed to sense and react to “friction” in emerge. An example is a security system that shares interfacing. If a thing encounters friction its key signifiers of trusted persons amongst components environment it can decide to use local adaptation of a without a central instance storing privacy-sensitive particular aspect that the thing embodies. This information. adaptation basically reduces friction in interoperation between the thing and its counterparts. Different from Discussion distributed cognition, this adaptation is a form of local The three categories of system properties as explained learning that each thing performs on its own. In above pose interesting opportunities for the design of systems of connected things, the creation, presence, connected things. Utilizing emergent or systemic decay, and absence of data determines the system’s properties of connected things is a function of (hyper-) 36 The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada connecting things, augmenting them with computation temporarily can be cumbersome and resource use and redesigning shared functionality with clear might be higher through duplication of services and “situatedness” and locality in mind. Locality refers to a resources locally. Finally, critical mass for emergent strong dependence and product relation to the local behavior is not always given. It is possible to imagine a context – in functionality and operational semantics. layer for individual people’s room in a shared Traditional designed products have naturally been local apartment, a layer for the apartment or house, and and self-contained (due to a lack of inherent then a layer for the apartment building or neighborhood connectivity, computation and data streams). In – each with a different purpose and functionality. But if contrast, most connected interactive products are there are too few things participating, no ecology of designed with a central controller, database or service things will develop. in place without which the product cannot or not operate as well. An emphasis on locality has benefits in Conclusion better control over privacy-sensitive data, protection In this position paper, we have introduced and against censorship, robustness against connectivity discussed a new frontier for designing IoT – the infrastructure problems, and the failure of single Intranet of Things. While many problems and components. Furthermore, we can adhere to share challenges of IoT systems and products are well- information instead of data, and quality instead of bulk. known, this concept is proposed as a new framing that Whenever new data is sensed and collected by is opposed to the conventional IoT vision that puts components of the system, they should make sense of cloud control, data leaking and privacy breaches first. T this data, enrich it with contextual “knowledge” and then reluctantly share it with other system components he Intranet of Things favors locality, decentralized on their request. This rule is central and applies to behaviors and creative use of communication pathways components or layers of a system, but also the system between things in a local context. Such communication as a hierarchical sub-system. Data trickles bottom-up, is meant to share data and to engage in reinforcing in gradually more processed and richer forms. behavior and thing-to-thing feedback loops. The However, even though we might be aware of possible creation of such new loops is an interesting field for properties and their opportunities, leveraging them in a design, creating awareness and giving access to design project is hard because technology is missing or leverage points that has not existed before. existing technologies have not been framing appropriately. Another problem might be that control of Future work decentralized functionality by a user might be more To support designers in creative applications of difficult or can only be done indirectly, maintaining a connected things, new technologies and technological consistent state or, and synchronization across different framings of existing technologies are required. Building locations requires again a connection. Processes that on such grounds, we need to develop frameworks for have established operating procedures such as creating systems prototypes using established design updating the system or shutting down operation software and hardware platforms that allow for fluent 37 The Future of IoT SmartObjects '18, in conjunction with CHI '18, Montreal, Canada use and rapid iterations through a new systems design Interaction Design. In Proceedings of the SIGCHI Conference on Human Factors in Computing process. We hope that through exploration like this Systems, 1053–1062. CHI ’09. New York, NY, USA: position paper and extensive discussions, a new ACM. doi:10.1145/1518701.1518862. generation of connected things will emerge that comply 7. Churchill, Elizabeth F. 2016. Designing Data to our intuitive understanding of things as meaningful Practices. interactions 23. New York, NY, USA: parts of the future Everyday. ACM: 20–21. doi:10.1145/2983401. 8. Williams, Meredydd, Jason R C Nurse, and Sadie References Creese. The Perfect Storm : The Privacy Paradox 1. Odom, William, Tom Jenkins, Kristina Andersen, and the Internet - of - Things: 1–9. Bill Gaver, James Pierce, Anna Vallgårda, Andy doi:10.1109/ARES.2016.25. Boucher, David Chatting, Janne van Kollenburg, 9. Perlow, Jason. 2016. All your IoT devices are and Kevin Lefeuvre. 2017. Crafting a Place for doomed. ZDNet. Attending to the Things of Design at CHI. interactions 25. New York, NY, USA: ACM: 52–57. 10. Abowd, Gregory D. 2012. What Next, Ubicomp?: doi:10.1145/3161605. Celebrating an Intellectual Disappearing Act. In Proceedings of the 2012 ACM Conference on 2. Svanæs, Dag. 2001. Context-Aware Technology: A Ubiquitous Computing, 31–40. UbiComp ’12. New Phenomenological Perspective. Human–Computer York, NY, USA: ACM. Interaction 16. Taylor & Francis: 379–400. doi:10.1145/2370216.2370222. doi:10.1207/S15327051HCI16234_17. 11. Ryan, William, Erik Stolterman, Heekyoung Jung, 3. Zhang, Ben, Nitesh Mor, John Kolb, Douglas S Martin Siegel, Tonya Thompson, and William R Chan, Ken Lutz, Eric Allman, John Wawrzynek, Hazlewood. 2009. Device Ecology Mapper: A Tool Edward A Lee, and John Kubiatowicz. 2015. The for Studying Users’ Ecosystems of Interactive Cloud is Not Enough: Saving IoT from the Cloud. Artifacts. In CHI ’09 Extended Abstracts on Human 7th USENIX Workshop on Hot Topics in Cloud Factors in Computing Systems, 4327–4332. CHI EA Computing, HotCloud ’15, Santa Clara, CA, USA, ’09. New York, NY, USA: ACM. July 6-7, 2015. doi:10.1145/1520340.1520661. 4. Rogers, Yvonne. 2006. Moving on from Weiser’s 12. Ryan, Alex. 2014. A Framework for Systemic Vision of Calm Computing: Engaging Ubicomp Design. FORMakademisk 7. Experiences. In Proceedings of the 8th doi:10.7577/formakademisk.787. International Conference on Ubiquitous Computing, 404–421. UbiComp’06. Berlin, Heidelberg: 13. Jung, Heekyoung, Erik Stolterman, Will Ryan, Springer-Verlag. doi:10.1007/11853565_24. Tonya Thompson, and Marty Siegel. 2008. Toward a Framework for Ecologies of Artifacts: How Are 5. Crabtree, Andy, and Peter Tolmie. 2016. A Day in Digital Artifacts Interconnected Within a Personal the Life of Things in the Home. In Proceedings of Life? In Proceedings of the 5th Nordic Conference the 19th ACM Conference on Computer-Supported on Human-computer Interaction: Building Bridges, Cooperative Work & Social Computing, 1738–1750. 201–210. NordiCHI ’08. New York, NY, USA: ACM. CSCW ’16. New York, NY, USA: ACM. doi:10.1145/1463160.1463182. doi:10.1145/2818048.2819954. 6. Odom, William, James Pierce, Erik Stolterman, and Eli Blevis. 2009. Understanding Why We Preserve Some Things and Discard Others in the Context of 38