
Industry executives and experts share their predictions for 2019. Read them in this 11th annual VMblog.com series exclusive.
Contributed by Sastry Malladi, CTO of FogHorn
Technology Takes Center Stage for Industrial and Commercial IoT Deployments
Recent research studies state the Internet of Things market
is expected to reach $1.6 trillion in industrial industries, $1.2 trillion in
commercial industries, and $1 trillion in consumer sectors within the next
couple of years. Moreover, companies see incredible improvements once they
translate this data, gathered by these connected machines, into actionable
insights using edge computing, machine learning (ML) and artificial
intelligence (AI).
Following the broader understanding and wider adoption by industrial
and commercial organizations of edge technology in 2018, technology supporting
IIoT will continue to evolve in 2019, including anticipated upcoming
developments such as edgification, closed-loop edge to cloud machine learning
and prescriptive maintenance will help lead companies to elevated efficiency
and cost savings.
Seven predictions to keep an eye on in 2019 are:
1. Survival of the smallest. IIoT
analytics and ML companies will be heavily measured on how much they can
deliver in how little compute.
As IIoT projects extend beyond
cloud-centric approaches, the next step in the evolution of artificial
intelligence and IIoT will address the need to convert algorithms to work at
the edge in a dramatically smaller footprint. According to Gartner, within the next four
years, 75% of enterprise-generated data will be processed at the edge (versus
the cloud), up from <10% today. The move to the edge will be driven not only
by the vast increase in data, but also the need for higher fidelity analysis,
lower latency requirements, security issues and huge cost advantages.
While the cloud is a good place to
store data and train machine learning models, it cannot deliver high fidelity
real-time streaming data analysis. In contrast, edge technology can analyze all
raw data and deliver the highest-fidelity analytics, and increase the
likelihood of detecting anomalies, enabling immediate reaction. A test of
success will be the amount of "power" or compute capability that can be
achieved in the smallest footprint possible.
2. The market understands
"real" versus "fake" edge solutions.
As with
all hot new technologies, the market has run away with the term "edge
computing" without clear boundaries around what it constitutes in IIoT
deployments. "Fake" edge solutions claim they can process data at the edge, but
really rely on sending data back to the cloud for batch or micro batch processing. When reading
about edge computing, the fakes are recognized as those without a complex event processor (CEP), which means
latency is higher and the data remains "dirty," making analytics much less
accurate and machine learning (ML) models are
significantly compromised.
"Real"
edge intelligence starts with a hyper-efficient CEP that cleanses, normalizes,
filters, contextualizes and aligns "dirty" or raw streaming industrial data as
it's produced. In addition, a "real" edge solution includes integrated ML and
AI capabilities, all embedded into the smallest (and largest) compute
footprints. The CEP function should enable real-time, actionable analytics
onsite at the industrial edge, with a user experience optimized for fast
remediation by operational technology (OT) personnel. It also prepares the data
for optimal ML/AI performance, generating the highest quality predictive
insights to drive asset performance and process improvements.
Real edge
intelligence can yield enormous cost savings, as well as improved efficiencies
and data insights for industrial organizations looking to embark on a true path
toward digital transformation.
3. ML/AI models get
skinny with edgification.
Moving machine learning (ML) to the
edge is not simply a matter of changing where the processing happens. The
majority of ML models in use today were designed with the assumption of cloud
computing capacity, run time and compute. Since these assumptions do not hold
true at the edge, ML models must be adapted for the new environment. In other
words, they need to be "edge-ified". In 2019, "real edge" solutions will enable
relocating the data pre- and post-processing from the ML models to a complex
event processor, shrinking them by up to 80% and enabling the models to be
pushed much closer to the data source. This process is called edgification,
which will drive adoption of more powerful edge computing and IIoT applications
overall.
4. Closed-loop
edge to cloud machine learning will become a true operational solution
As machine learning (ML) and AI algorithms become "edgified"
for use close to sensors or within IoT gateways or other industrial compute
options, new best practices will emerge on how to train and further iterate on
these models. What industrial organizations will find is that edge devices
generating analytics on live streaming data (including audio and video) should
regularly send insights back to the cloud, but only those that represent
anomalous activity warranting a shift in the core algorithms. These edge
insights enhance the model, significantly improving its predictive
capabilities. The tuned models are then pushed back to the end in a constant
closed loop, reacting quickly to changing conditions and specifications, and
generating much higher quality predictive insights to improve asset performance
and process improvements.
5. Production
IIoT applications will go into implementation only with edge computing
solutions supporting multi- and hybrid-cloud deployments.
Hybrid- and multi-cloud solutions
will dominate the industrial IIoT deployments - a recent report found that the
hybrid-cloud market will reach $97.64B USD by 2023. As industrial organizations
look to bring multi-cloud environments together to provide a more cost
effective approaches and flexibility, it will be important for edge solutions
to be cloud agnostic. Vendor-exclusive solutions will likely begin to fall by
the wayside as companies look for more flexibility and freedom of choice when
building their edge-to-cloud environments. Google, AWS, Microsoft, C3IoT,
Uptake and other leading cloud providers will establish more collaborative
partnerships with edge computing companies to help businesses as they continue
to improve and expand their offerings.
6. IoT video and audio sensors take off,
driving the need for deep learning at the edge.
There is industry-wide excitement
about the capabilities that audio and video sensors can bring to the IIoT. Edge
computing technology can play an important role in the further deployment of
audio and video data in commercial and industrial IoT systems. The fusing of
asset data with audio and video analytics will allow for faster and more
accurate device and machine maintenance (including updates on systems health
and more), and a whole host of new innovative applications. One such example of
the video analytics is the use of flare monitoring at oil and gas operations to
track environmental compliance and flare state remotely for large volumes of
flare stack towers.
7. Predictive maintenance gives way to
prescriptive maintenance.
One of the big promises IIoT edge
solutions deliver is predictive maintenance, offering insight into what is
likely to happen to a connected asset (like manufacturing equipment or an oil
rig) in the future. While many organizations still lag in implementing
predictive maintenance as a first step, even more advanced technology will be
available to early adopters in 2019.
Prescriptive maintenance is a step
forward to enable businesses to not only predict problems, but also produce
outcome-focused recommendations for operations and maintenance using data
analytics.
For example, elevator manufacturers
want to put an end to routine problems, such as friction in elevator doors. As
part of this effort, they partner with Foghorn to create a predictive
maintenance solution. By analyzing sensor data at the source, they can now
determine maintenance needs well in advance, without the cost, latency,
security and other issues associated with transfer of large amounts of data
outside of the building. Thus, it can schedule service before anomalies impact
performance in a highly efficient manner. As prescriptive maintenance becomes
available, before the manufacturers roll a truck to provide maintenance on an
elevator, they will have data available to suggest areas most likely to need
repairs and have verified the repair staff person the expertise, tools and
parts available for the repair.
##
About the Author
As CTO of FogHorn, Sastry Malladi is responsible for and oversees
all technology and product development. Sastry is a results-driven technology
executive with deep technology and management experience of over two and half
decades. His areas of expertise include developing, leading and architecting
various highly scalable and distributed systems, in the areas of Big Data, SOA,
Micro Services Architecture, Application Servers, Java/J2EE/Web Services
middleware and cloud Computing to name a few.
Prior to joining FogHorn as CTO, Sastry was Chief Architect
of StubHub, an eBay company where he led the technology architecture
transformation and also spearheaded the Big Data initiatives and data driven
decisions. Sastry was also a key technology executive at eBay that lead the
technology re-platforming effort from its monolithic architecture to the
distributed, and scalable service-oriented architecture that it is today
enabling the business growth. Prior to joining eBay, Sastry was co-founder and
CTO of OpenGridSolutions, Founding member and Architect at SpikeSource and an
Architect at Oracle.