Cloud and IoT – which to choose?

The market of cloud providers with IoT services is developing and maturing fast. Cooperation and eventually consolidation are happening continuously with frequent announcements of alliances and/or acquisitions of smaller providers.
Figure 21 presents Forrester Research’s view of current “evaluated vendors against 24 criteria, which we grouped into three high-level categories:

Current offering. Each vendor’s position on the vertical axis of the Forrester Wave graphic indicates the strength of its current offering. Key criteria for these solutions include connectivity, deployment and security, the management console, application enablement functions, and analytics and data.
Strategy. Placement on the horizontal axis indicates the strength of the vendors’ strategies. We evaluated partner strategy, commercial model, innovation roadmap, platform differentiation, and delivery model.
Market presence. Represented by the size of the markers on the graphic, our market presence scores reflect each vendor’s direct customers, connected devices, and geographic customer distribution.”

Cooperation & Consolidation

Trending is cloud providers teaming up to dominate business verticals based on the original characteristics of each supplier.
For instance, General Electric is cooperating with Microsoft to combine GE’s digital transformation consultancy on industrial equipment with Microsoft Azure database & BI platform and the ability to deploy infrastructure remotely.
Cisco traditionally delivers networks infrastructure & connectivity components and combined with IBM’s backend and Watson advanced analytics they complement each their specialties.
For several hospitals Cisco and GE join forces to improve patient and asset optimisation to reduce waste of displaced articles and dormant or non-optimised use of equipment – cooperation between the large suppliers is mixed and matched across industries and verticals.
PTC (lifecycle management SW) acquired ThingWorx together with Axeda, both individually adding experience from industrial IoT and bringing in a component platform of easy mashup and visualisation with statistical components respectively. Later ColdLight Machine Learning and predictive analytics has been added to the portfolio.
In comparison Amazon or Google, whose services tend to be delivered solely by their public clouds, offer minor experience with physical devices on the edge. Recent collaboration between Amazon and Salesforce has been announced.

Selection criteria examples

Selecting a commercial cloud service requires a thorough analysis of the demands and priorities of the business to be supported. The following is a sample list of criteria to consider:
Geography, local instances globally: Local legislation may demand personal data retained inside domestic data storage only. For other data types faster response rates and data redundancy are strengthened by cloud storage localised on the same continent.
Flexibility between Cloud and the Edge: Protocol support for both short-range and wireless connectivity options. The ability to deploy a ‘hybrid’ or ‘local’ cloud, i.e. perform time critical calculations on computational units closer to the devices to guarantee latency and stability (e.g. off-shore wind turbines or cargo ships). Further, for scarcely connected Industrial IoT plants minimise the risk of data-loss if data transfer is disrupted. Reduce transfer load if the cost of data is high.
Development and experiments tools support: Software Development Kit (SDK) for modelling fast data analytics and drag-and-drop visualisation with user interaction modules. Otherwise will low level coding be necessary?
Advanced analytics & AI: Those terms cover across a vast span; from standard statistical analytics to automatic (supervised or unsupervised) pattern recognition and learning algorithms building confidence on still larger sets of apparently uncorrelated data. Certain cloud providers assign almost human traits to their learning capabilities.
Scalability: How does the supplier treat companies with simple sand-boxed experimental data sets; how to increase load and support when rolling out broader deployment; and how to cap cost and branding large scale operations globally?
One-stop shopping and support: The ability to offer support and mitigate problems across external territories. Will the supplier proactively assist in finding errors outside their own specific business domain?
Integration with production Master Data: The ability to link to (e.g. SAP) Master Data from the production logistics for integration with lifecycle management of each product manufactured represented by digital twin functionality.
API management: Interfacing to external sources (in/out) of e.g. weather data, map data aligned with fleet management, mashup of specific data sets from 3rd parties or supporting a platform for 3rd party data brokering.
Business models & cost structure: Basically all of the above characteristics and functionality may impact the price model of the cloud supplier: data collected (sample rate, number of data items, average/peak bandwidth, etc.), total storage space, background (cold data) analysis, real-time (hot data) analysis, complexity of analysis & presentation needs, deployment and redundancies across geography, development support versus tools support, etc.

How does Beacon Tower fit?

As you probably have already read elsewhere on this website, Beacon Tower is a very flexible IoT platform that is utilising the many managed services supplied by the Microsoft Azure cloud combined into a complete package with respects to deployment, monitoring and maintenance. So, in comparison to the above selection criteria and on top of the advantages that Forrester points out for Microsoft Azure, Beacon Tower is a an even more accessible, flexible and powerful IoT solution than what you get ”out-of-the-box” when trying to design your own solution with standard Microsoft Azure components. Read more about the platform here.

Continue reading: go to Eco-systems: Make, Buy and Partner