Introduction
In a previous research entitled “The Vital Role of Edge Computing for IoT: 2016 Update“, Wikibon laid out the IoT cloud imperative that almost all IoT should be strategically implemented using an IoT Hybrid Cloud Model. It concluded that most data (95%+) in IoT will be created, processed, stored and analyzed and be discarded at the Edge. The research illustrated the business cases with multiple scenarios, and showed that the Edge Hybrid IoT Cloud Model was by far the most cost effective. In addition, all the technology winds were increasing the potential benefits of Edge IoT processing.
The technology conclusions of the previous research are that the biggest challenges for IoT are creating the distributed networks, distributed databases and distributed analytic software that can be managed from a site remote from the Edge, without any local expertise required. In addition, the design of networks must allow clear air-gaps and enable end-to-end encryption. The IoT Edge must have the ability to create and tear down distributed connections with other parties, to make and verify transactions with limited data sharing with other parties instantaneously and without a central authority. Blockchain technologies are a good starting point for the design of such systems.
This research is focused on evaluating the partner options available to IT executives for implement such an approach. The IoT solutions are generally in an early state of development, and the leadership issues are far from resolved. This research will be updated as significant changes are made.
It should be noted, as in the research referenced above, that traditional hybrid clouds are very different from the IoT Hybrid Cloud. This research does not seek to evaluate all models of hybrid cloud, only IoT hybrid cloud.
Partnering for a Hybrid Model of IoT
The major public cloud vendors in the US are Amazon, Google, IBM and Microsoft. Microsoft has announced its own hybrid strategy with Azure stack. GE has a well-developed hybrid Predix system both distributed and cloud based. HPE have a Universal IoT Platform, with a significant strategy to “shift left”, putting more functionality towards the Edge. Oracle has an aggressive private/public cloud hybrid strategy and is strong in distributed database functionality. VMware and AWS have announced a hybrid-light strategy.
This research will analyze the potential partners for implementing a Hybrid Model of IoT.
Amazon as an IoT Hybrid Model Partner
In 2015, Amazon announced a centralized IoT set of services on AWS that allows full management of IoT message streams in the AWS public cloud. AWS is integrated with AT&T NetBond® to enable network connections to the AWS Cloud. Recently, the two companies expanded the agreement and will jointly develop solutions that span cloud networking, mobility, Internet of Things (IoT), security and analytics.
Amazon has also recently announced a partnership with VMware to host VMware on bare metal AWS servers, with a close integration with AWS APIs. The advantages of this approach for enterprise IT are:
- Low cost VMware IaaS service
- Very low latency integration between AWS cloud services and VMware. VMware technologies such as vMotion allow applications and data to be moved to the cloud, and be temporally close to applications running on AWS.
This allows AWS a hybrid strategy once-removed by partnering with VMware for the public cloud and distributed VxRail systems to run IoT software at the Edge. Pivot3 runs under VMware and could be included in such a strategy. The AWS software has little to offer in these VMware-driven hybrid scenarios.
The solution necessary for AWS to compete in this IoT Hybrid Cloud model long term area is to develop distributed AWS software packages that can be run on distributed systems at the Edge, and allow them to be managed from the AWS cloud. AWS’s customers have been forthright in asking for such a system.
Wikibon assessment of AWS for Enterprise Hybrid IoT: AWS supports a “bring-your-own-system” hybrid model, with a set of APIs to help manage the two environments. However, AWS does offer distributed databases, distributed time-series databases, distributed analytics, Blockchain technologies, or complete remote management of the IoT hardware and software. Consider using AWS for initial projects for speed-to-test initial solutions. Until AWS invest significant resources and products in an IoT Hybrid model, avoid authorizing large-scale deployment on AWS.
AWS remains a formidable low cost Infrastructure-as-a-Service (IaaS) vendor, with the ability to host either marketplace IoT Cloud software or SaaS vendors that manage the interface with Edge devices and data with an option of using network services from AT&T NetBond or others.
GE as an IoT Hybrid Model Partner
Wikibon has reviewed GE’s approach to industrial IoT, based on their IIoT platform called Predix. Predix takes a hybrid architected approach to both support the edge needs and exploit the scale and analytic choice of the cloud. GE’s own experiment has started with the premise that systems must meet the requirement of both the IT and OT communities.
At the heart of Predix is a connectivity architecture that connects sensors, control systems, gateways, and equipment management needs typical of OT, to the appliance, server, and storage needs to process that data. A multitude of OT and IT protocols are supported.
Predix provides a distributed messaging model to handle alarms and events, time series databases, and analytic data needs. Predix provides a set of “Industrial Micro-services” for asset, analytics, data management, security, and operations functions. Predix also offers a set of components for handling edge computing. These include a engines for rules, expression, and aggregation, as well as data handling components, local database services, and the ability to synchronize the 5% or less of data to the GE cloud.
Wikibon assessment of GE Predix for Enterprise Hybrid IoT: GE is both a practitioner and developer of IoT solutions. GE’s Predix offers a hybrid approach to Industrial IoT, and one that can facilitate the discussions between OT and IT. There is still room for development of more distributed time-series databases,remote management and other distributed services. GE partner with HPE, Pivot3 and other vendors for Edge hardware and software.
Google as an IoT Hybrid Model Partner
“Brillo” is Google’s Edge IoT operating system deployed on the Android kernel. Google’s “Weave” is an open communications platform for IoT, and by integrating with Google services (including analytics) provides the decision making/advising based on IoT data. Google has similar philosophy to AWS and is public cloud centric, with all the data being processed in the cloud.
Wikibon assessment of Google for Enterprise Hybrid IoT: Google does have an IoT operating system that will run on low cost Android hardware. However, Google does not support a hybrid model, in the sense that data can be mainly processed all the Edge. Consider using Google for initial projects for cost reasons, but until Google invest significant resources in a hybrid mode, avoid authorizing large-scale deployment on Google.
HPE as an IoT Hybrid Model Partner
HPE is a leader system vendor with broad range of offerings, including expertise in ARM processors which is required for connection to the sensor ecosystem. The HPE Universal IoT platform gives an end-to-end architecture with a single throat to choke for private cloud and distributed IoT. HPE works with the Microsoft Azure Public cloud as a primary partner. HPE expertise in vertical IoT software is not so strong, and relies on HPE partners.
Wikibon assessment of HPE for Enterprise Hybrid IoT: HPE fully embraces the IoT hybrid model, and has a full range of hardware and system software. HPE should be considered as a potential partner as part of a consortium with other partners, as for instance the GE Predix software discussed above.
IBM as an IoT Hybrid Model Partner
IBM is a strong software company, with a strong services offerings. IBM has deep expertise in the manufacturing space with both distributed hardware and software packages. However, IBM has announced it will be “deprecating” (obsoleting) the Time Series Database service on BlueMix; time-series a key component of IoT. IBM has excellent analytic and AI software such as Watson. However, IBM is focusing on Watson as a SaaS service, which will not allow distributed versions at the Edge. IBM also has excellent consultative services in many areas, including IoT. However, IBM’s inability to deliver hybrid IoT software as part of a solution makes any partnership with IBM of limited value.
Wikibon assessment of IBM for Enterprise Hybrid IoT: IBM does not embrace the IoT hybrid model, and has does not have a full range of hardware and system software. Wikibon does not recommend IBM as an IoT partner.
Microsoft as an IoT Hybrid Model Partner
Microsoft has an Azure IOT Suite as well as the Microsoft Azure Stack as IoT building blocks. Microsoft have put an emphasis on security, including network segmentation, the need for isolation (air gap) solutions and hybrid cloud-based security. Microsoft have utilized Microsoft Active Directory and the recent announcement of Microsoft distributed databases as key components in IoT. Microsoft are new in industrial IoT, and will need to partner to provide OT and IT skills by vertical.
Wikibon assessment of Microsoft for Enterprise Hybrid IoT: Microsoft Azure stack is a true hybrid implementation, with a solid range of system software, middleware, security and application software. However, Microsoft are new to IoT, with limited understanding of manufacturing. Microsoft should be considered as a good potential partner with a strong commitment to a compete hybrid cloud IoT architecture. They should be considered as part of a consortium with other partners, as for instance with the GE Predix software discussed above, with HPE or Dell for hardware, etc
VMware as an IoT Hybrid Model Partner
VMware has little experience of manufacturing or other IoT environments. VMware has a rudimentary IoT offering using open source software called Liota.
Wikibon assessment of VMware for Enterprise Hybrid IoT: VMware is very early in the game and has almost no presence in IoT, either alone or in the Amazon Cloud. Enterprise IT should look for vendors with solid offerings and partnerships that use VMware (e.g, Pivot3).
Conclusions
Action Item:
Enterprise executives with responsibility for implementing a long-term strategy for IoT should ensure that the architecture they adopt is a full hybrid cloud solution, capable of storing, processing and deleting 95% of data at the Edge today, and managing 99% of data at the Edge by the end of the strategic project. Wikibon recommends that initial Hybrid IoT discussions be with GE, HPE and Microsoft, with AWS as a potential IaaS supplier.
