eWeekChat

A4: By making edge computing more available, easier to use, and better integrated with cloud, it becomes more enticing to developers and more worth their time to learn about.

A4: As scalability, interop and connectivity are slowly figured out, the IT/OT divide aka “on-prem silos” needs to be solved with more upfront tooling to make buildings and infrastructure smart out-of-the-box. IT needs to take the lead on harmonizing the infrastructure for IT/OT

A4 - In order to address these challenges, it is important that companies interested in realizing the benefits of fully local edge computing look for partners whose solutions have been architected to work at the edge from the ground up.

A4: (1/2) Some of these are the natural outcome of the relative ‘newness’ of #edgecomputing & will resolve over-time. Establishing a longer-term strategy for edge, while allowing near-term ‘start small/pilot projects’ will also partially mitigate many of the challenges.

A4. Planning is crucial. Companies must identify and develop the right use cases and proper integration strategies, including proactive security and #privacy-first approaches. Even with the best edge computing plan, ignoring #security issues can be a fatal mistake.

A4: The edge needs onsite IT infrastructure that is low cost, easy, quick access to data and AI.

@RobertBlumofe Do you a problem with adoption if edge services lag behind cloud native services?

A4 - First understanding what you need to solve for in terns of latency, security, and reliability. Matching and pivoting your application stack and technologies to meet the needs. The space is maturing quickly.

A4: (2/2) Additionally, re-using components already in-place for #cloud, such as management & #orchestration tools, can simplify the complexity and help scale challenges.

@BMCSoftware Yes, cloud and edge integration!

A4 -- AWS is pretty darn distributed now, it's not east and west coast US and limited to a few countries. Experiment?

@brucekornfeld what about the skills to maintain it?

@enlightedinc I don't think the problem is so much in lag as it is integration. Edge and cloud need to be integrated so they work together to solve a larger problem.

@nordmark_erik made a good point in another post -- portability, if edge infrastructure is heterogenous (it is) having common OS / app stack is key

A4: I think companies need to try it out and learn by deploying something from the cloud at the edge.

A4: But first they need to understand what is unique at their edge; retail might be very different than oil and gas due to different "remoteness".

A3: Mainly the challenge is in educating developers about how and when to use edge computing. Complexity is still a challenge but edge computing platforms are becoming easier to use and better integrated with cloud.

A3a. While distributed IT environments inherently create #complexity, #security and #privacy present huge challenges. Edge expands the risk surface & chances of #ransomware attacks.

A3b. Other challenges would be a proliferation of technology, a lack of standardization, and shortage of skilled personnel. With proper planning, organizations can implement solutions that address these challenges.

A3: We see the reverse challenge of the “IT/OT” divide (aka “on-premise technology silos”). Many of our core users still opt for fully on-premise over cloud-connected. IT often leaves “operational technology” (e.g. building systems) siloed from their strategies.

A3: Definitely security and complexity. They need to make do with no or limited staff. But also not realizing what is hard and what is easy with edge deployments.

@prasadjoshi_atl Excellent point. Security has to be factored in.

A3: (1/2) #Edgecomputing remains fragmented & verticalized with a lack of consistent standards that allow solutions to span multiple business siloes & address the main challenges of scale, edge connectivity & #data privacy in a consistent manner.

A3 - its a new model, failure modes are / can be different, cell networks and topology are different than typical static IT, security risks may be higher, costs may be higher......

A3: (2/2) This is compounded by poor understanding of edge computing by decision leaders & tactical purchasing of point solutions.

A3: The biggest edge challenge we see is delivering increasingly more advanced #AI applications running on constrained #smarthome devices (CPU, memory, etc.) that can act autonomously of a constant network connection.

@nordmark_erik There's not a real homogenous platform that can meet all location needs and platform needs -- still very much a work in progress as a platform.

A3 - The cloud is one of the problems. Over the last decade or so, most IT teams have figured out how to leverage cloud - but they are learning that in many cases, cloud is too expensive, slow and unreliable for the real-time compute, storage and analysis needed at the edge

A3: Deploying e.g., kubernetes running in the cloud at the edge seems easy, but 80% of running it is the infrastructure which is already there in the cloud (OS updates, physical security, reliable network)

@jtcarolan Yes, there are important architecture considerations around async communications and data updates.

@jtcarolan right! and this makes it hard for users to figure out the right architecture for them - too many choices!

@jtcarolan we are trying to improve the edge part of the platform ecosystem by opensourcing EVE-OS under the Linux Foundation.

@prasadjoshi_atl Return of the "server room" ???? What about the server under the office desk to service a retail environment?

@nordmark_erik https://www.lfedge.org/projects/eve/

A6. Edge spending will continue to increase, as we see other industries like #automotive, #retail and #healthcare increasingly relying on edge. We’ll also see #6G rollout boost edge computing capabilities, with the possibility of space exploration providing even more momentum.

A6: A maturing market increasingly orientated around #hyperscaler delivered extensible edge platforms, with vertical specific capabilities added by industry partners to meet greater customer demand for holistic, horizontal solutions focused on scale, cost efficiency & security.

A6: Edge computing will enable new generation of IoT devices that are more secure & lower cost. By moving functionality from device to edge, the device is effectively hollowed out with the on-device software only transducing from the physical world to cyber.

A6: there will be Standard applications and architectures, data analytics packages that can be tailored to use-cases, easy to build data retention/movement policies,

@brucekornfeld Probably not enough.

A6: 1/ Big tech will dominate edge computing infra, but industrial OEMs and startups will be the ones to solve for specific use cases. 2/ AI/ML model lifecycle management will be possible fully at the edge. 3/ Volume of data transferred from edge to cloud will decrease relatively

A6 - standard will emerge, likely lead by the hyperscalers, people will talk more about actual use cases, developers will understand latency, its impact, and distributed topologies better. "Good enough" will always be cheaper tho for many use cases.

A6 - AI will also be an accelerant here -- more IoT data sources, edgier nodes that can receive/process/store/forward -- more data for AI to do more use cases.

A6 - There will be an explosion of devices that are able to run #AI at the edge. These devices and apps will be able to shift some computing to clouds that can provide extremely low-latency resources, over #5G, for example.

A6: Being slightly contrarian, I think enterprises will realize the value of their data and want to retain full control of it. They will continue to use the cloud and the edge, but depend less on any particular hyperscalar.

A6 - There will be an explosion of #IoT and #smarthome devices that are able to run #AI at the edge. These devices and apps will be able to shift some computing to clouds that can provide extremely low-latency resources, over #5G, for example.

A6 - There will be an explosion of #IoT and #smarthome devices that are able to run #AI at the edge. These devices and apps will be able to shift some computing to clouds that can provide extremely low-latency resources, over #5G, for example.

A6 - There will be an explosion of #IoT and #smarthome devices that are able to run #AI at the edge. These devices and apps will be able to shift some computing to clouds that can provide extremely low-latency resources, over #5G, for example.

@nordmark_erik good point, it will be a mixed bag, hyperscalers will support the "ownership" as well

@OriginWireless AI at the edge. Yes!

A9. Enterprise edge computing will be inevitable in the future, as users demand a better experience, more engagement, and businesses rely on real-time analytics. With this shift, enterprises should start creating a plan now to innovate & implement #edge across their business.

A9: Edge computing will play a pivotal role in the next generation of AI Large Language Models and Machine Learning. Running LLM and ML models at the edge will enable the requisite scale and low latency.

A9 - we've been in a highly centralized architecture model for years -- "mainframe" client server etc. Experiment with app service layers and data layers that help reduce latency on highly sensitive transactions or use cases.. near edge, user edge --- relative.

A9: A lot of what they do in the cloud applies to the edge, but need to understand what is different whether that is ruggedized hardware, physical security. network outages.

A9: The #enterprise edge is a diverse tech capabilities. While new solutions are required to meet edge needs, growth in ‘cloud-out’ & #hyperscaler edge computing means existing management & orchestration tools remain important to the edge toolkit.

A9: Edge Computing isn't new. In the Mainframe days, remote sites connected to "the brains" with dumb terminals. Technology has evolved so now these remote enterprise (EDGE) sites don't have to be dumb - CPUs are so inexpensive and powerful, data storage is advanced.

A9: The trend behind edge is inevitable as a way of increasing efficiency. More complex problems will be solvable closer to where the problem is happening (using LMMs, GANs, etc..). Organizations with great strategies around edge will outperform those without!

A9: (con't) They should consider all possible ways to solve their IT challenges - cloud solves some, but a robust, onsite edge computing solution can solve many as wel..

@brucekornfeld Yup, they solve different challenges and work best together.

A9: Look for commonality across projects and organizations otherwise you might end up with a mess of edge strategy with every project doing it differently for no good reason.

A9: For #IOT and #SmartHome solutions, edge computing enables the privacy, security, performance, and resilience that are foundational to delivering great customer experiences.