In software engineering for the last 15 years - Lead engineering at Tyk.io - Happy father living a nomad life https://github.com/buger https://speakerdeck.com/buger @buger [email protected]
on the centralized, server/client paradigm to authenticate, authorize and connect different nodes in a network. IoT going grow to hundreds of billions of devices. Centralized systems turn into a bottleneck.
on the centralized, server/client paradigm to authenticate, authorize and connect different nodes in a network. IoT going grow to hundreds of billions of devices. Centralized systems turn into a bottleneck.
technologies competing to become the standard. Some will become obsolete in the next few years, effectively rendering the devices implementing them useless. Physical device longevity now depends on your software
Machines’ actions in unpredictable situations - Information security and privacy - Machine interoperability - Mean-reverting human behaviors - Slow adoption of new technologies
requirements change constantly - New uses for devices - The uses for Internet of Things technology are expanding and changing—often in uncharted waters. - Consumer Confidence - Lack of understanding or education by consumers of best practices for IoT devices security (default passwords?)
for data markets are missing especially for data brokers: how data is used and who has access - Even though data appear to be the currency of the IoT, there is a lack of transparency about it
(M2M)/"Internet of Things" connectivity protocol - Lightweight, open, simple, and designed for easy implementation - Small code footprint where network bandwidth is at a premium - Ideal for constrained communication environments - Continuous session awareness (publish a message if client goes offline)
client subscriptions on topics - Receive messages from clients - Forward messages to interested clients MQTT Broker - Publish messages to the broker for other clients to receive - Subscribe to topics they want to receive MQTT Client
ClientID Or Broker/Gateway assign ID Each client gets own token, with base policy, which you can trace via analytics, and individually configure granular ACL.
rewrites to re-route topics - Body rewrites to modify payloads - Conditional logic based on metadata (like QoS) - Load balancing and service discovery - Configuring QoS, retain or last will logic - Internal tooling like ElasticSearch or StatsD
Thousands of clients Standard 2 Core cloud VM instance: 8 core dedicated machine: - 80k RPS - Tens of thousands of clients - Few ms added latency with auth and analytics