Sunday, March 31, 2019

AI, IoT, Edge Computing, 5G are Part of an Ecosystem

Ultimately, it might be difficult to clearly separate applied artificial intelligence from the value of internet of things, from edge computing, from 5G. The value of IoT sensors often will be real-time data--and lots of it--analyzed quickly, on the spot, to produce insights that in turn allow modification of behavior.

Companies are finding that machine learning can have significant advantages over traditional business intelligence tools for analyzing IoT data, including being able to make operational predictions up to 20 times earlier and with greater accuracy than threshold-based monitoring systems, argue consultants at Deloitte.


Saturday, March 30, 2019

Schneider Electric Touts its Micro Data Centers

MetroEdge Goes for a Niche

There are niches, and there are niches. That will be true for the edge computing ecosystem as much as it is true for other industry segments. Consider MetroEdge which wants to build a national grid of “state-of-the-art, high-performance, micro-scalable data center computing facilities in traditionally underserved urban/metro areas.”

“MetroEdge is a Minority Business Enterprise (MBE),” the company says. As has proven to be the case in the past, that strategy entails getting a percentage of larger contracts won by larger firms, since the federal government often has requirements or preferences that some percentage of new contracts go to firms headed by people of color or women.

The business strategy is arbitrage of government contracting rules, and there is a proven niche strategy at work, at least for small firms.

“Our high performance Edge Computing mesh network of micro data centers in urban areas across the U.S. generates significant economic development in underserved communities, while enabling low-latency data transmission needs for the autonomous vehicle industry, IoT, and local Fortune 1000 business demand for high-compute private cloud services, while leveraging investment in QOZ’s (qualified opportunity zone’s),” Metro Edge says.



Friday, March 29, 2019

Edge Computing Economics Driven, in Part, by Sheer Processing Volume

Sometimes quantity becomes qualitative change. Perhaps the classic example is the change of state of water as it is heated or cooled. At some point, a small accretion of temperature changes results in a quantum change of structure.

In the same way, the sheer volume of internet of things data to be analyzed will create new imperatives for analysis and bandwidth.

By 2022, 75 percent of enterprise-generated data will be created and processed outside the traditional centralized data center or cloud, up from less than 10 percent in 2018, according to Gartner.

Even when speed of response is not an issue, it will make more sense to process data locally   rather than sending it across the wide area network, as it will save operating expense.

IDC research predicts 45 percent of IoT-created data will be stored, processed, analyzed, and acted upon close to, or at the edge of, the network within three years (2022 or so)  and over six billion devices will be connected to edge computing solutions.


The point is that the increase in IoT sensor data, especially bandwidth-intensive data, will create new economics for local processing and analysis, compared to shipping data out to the cloud for analysis.   

The business case for local processing (on-site enterprise or edge data centers) will get better as the volume of data to be analyzed grows exponentially. So even if ultra-low latency remains the unique driver for edge computing, the practical concerns of WAN bandwidth charges will create equally-strong financial drivers for local processing.



50% of Industrial IoT Operations at the Edge by 2022

Enterprise edge computing could represent half of all industrial internet of things analytics workload by 2022, Stratus believes. That might represent a virtually zero revenue opportunity for other forms of edge computing conducted in the local area, but not on the premises.

One obvious use case for off-the-premises edge computing are analytics tasks supporting sites too small to use a fault-tolerant local computing solution. Other use cases might include analytics supporting widely scattered or moving sensors, as well as support for local analytics for smaller businesses and organizations that cannot justify the cost of a premises capability.

Wednesday, March 27, 2019

75% of Enterprise Data Created and Processed at the Edge?

By 2022, 75 percent of enterprise-generated data will be created and processed outside the traditional centralized data center or cloud, up from less than 10 percent in 2018, according to Gartner. That is why some are sure edge computing will be a “thing.”

source: HPE

How Big is Fog Computing in 2022?

You would expect an organization such as the Open Fog Consortium to have a robust belief in the growth of fog computing revenues, especially in industrial IoT settings. The issue is how much of that revenue might accrue to connectivity providers.  


Precisely where the edge computing happens--and what we call it--varies. Some prefer the term “fog,” while others use “aggregation edge,” “premises edge” and “device edge” to describe different logical places edge computing can happen.
source: Open Fog Consortium

New Tools for Edge IoT by Cloudera



Cloudera DataFlow (CDF) is an edge-to-enterprise real-time streaming data platform that could have application for internet of things sensor deployments.  

Sunday, March 24, 2019

Will 5G, Edge Computing Help Create the Next Era?

Will edge computing and 5G be part of the next big shift of communications service provider business models? Many now believe so. And one big implication might be that communication networks matter less than they used to, in terms of creating the value of “communications.”

Yes, networks are fundamental. But so is electricity, and it is hard to say that 20th century or early 21st century innovations are created directly by electricity. Rather, it is all the things made possible because electricity is available that matter.

That is not to belittle or minimize the centrality of electricity. Try and imagine living without it for an extended period of time. Still, electricity is a foundation for the devices, software, platforms, use cases and solutions that depend on electricity.

And we might be approaching an era where communications capabilities increasingly resemble electricity. Perhaps it always has been so.          

Technology is “the application of scientific knowledge for practical purposes.” The key notion there is of tools created for some human purpose. If you have been in the technology or communications business long enough, you are accustomed to hearing about “solutions” rather than software, hardware or technology as the “product” being sold by a particular supplier.

Nokia Bell Labs has positioned the future of networking as one where the value proposition for connectivity is quite different. The thesis is that communications network value will be created to the extent that it “creates time” for people and augments human intelligence.

Nokia does that mean that in the astrophysics sense of manufacturing time itself. Think of the word “productivity” or “efficiency” and the sense arguably is clearer. “Save you time” is probably the best way to think about the broad value proposition for communications networks of the future.

People will still want to communicate, share information, thoughts and ideas. But there will be a shift towards modes that allow richer conveyance of emotion and feelings, Holt suggests. That was touted as a key advantage of “telepresence” a decade ago.


In other words, connectivity is not going to be the reason businesses, organizations and people spend money on communication capabilities, in the sense of “people want holes, even when they buy shovels.”

That is akin to the past thinking we have used about the value of any computing or communications capability: it is the solution to a business or consumer problem, not an end in itself. People do not spend money on smartphones because they like carrying little computers in their purses and pockets.

The devices only enable the conversations they can have with people; the information they can discover immediately; the transactions they can conduct; the music and video they can watch. It is likely that sense which drives the Bell Labs notion of value driven by the ability to “create time.”

Nor would it be in keeping with the concept to pin all the coming changes on 5G or communications capabilities, but on a complex of changes anchored by artificial intelligence, which Bell Labs prefers to call “augmented intelligence.”
source: Nokia Bell Labs  

Wednesday, March 20, 2019

Google Stadia: Does it Compete with Edge Computing?



Google argues that Stadia, its new platform for cloud-based gaming, will perform as well as locally-stored apps. Which of course raises the question: if that is true, then maybe edge computing is not really needed. But there is skepticism. 

Tuesday, March 19, 2019

Cloud Kings Move to the Enterprise Edge

Today’s cloud kings already are moving to support enterprise edge computing running on the same tools enterprises would use when buying public cloud computing services. And that is intended to make today’s cloud kings big players in hybrid cloud or multi-cloud computing.

Those moves might also mean the cloud kings could emerge as major players in enterprise edge (on the premises) forms of edge computing.

“Latency—the time it takes for data to reach a hyperscale data center and for answers to return—is too much in industries such as retail, factory manufacturing, and financial services,” said IDC Chief Analyst Frank Gens

As a result, he said, the last 12 months have seen an explosion of edge and on-premise offerings by the cloud kings such as Azure Stack , AWS Outposts and Greengrass (the AWS IoT edge solution) , IBM Cloud Private, and Google Cloud IoT Edge.

“AWS Outposts bring native AWS services, infrastructure, and operating models to virtually any data center, co-location space, or on-premises facility,” AWS says. “AWS IoT Greengrass seamlessly extends AWS to edge devices so they can act locally on the data they generate, while still using the cloud for management, analytics, and durable storage.”

Google Cloud IoT is a complete set of tools to connect, process, store, and analyze data both at the edge and in the cloud. The platform consists of scalable, fully-managed cloud services; an integrated software stack for edge/on-premises computing with machine learning capabilities for all your IoT needs.

"These are becoming the stacks that matter," he said, noting that IDC predicts that by 2023, 30 percent of all the IT systems in enterprise data centers and edge locations will be running public cloud-sources services.

In other words, the enterprise edge (on-premises computing) will largely be fueled by the same firms that dominate today’s cloud computing business.

Sunday, March 17, 2019

Fairly Obvious Why Server Suppliers Believe in Edge Computing



Lots of servers would need to be put out onto the edge. Dell EMC iVP Jimmy Pike speaks about what Dell EMC is preparing for the market.

Friday, March 15, 2019