eWeekTweetchat

Future of Edge Computing
Experts discuss the future of edge computing.
James Maguire
Q8. Big Question #2 of 2: What do you see as the long term future of edge computing, say a decade plus?
Charlie Ashton
A8: That for out, edge computing will operate like a utility delivered via a grid, where cloud computing resources are available when and where they are needed, under a pay-as-you-go consumption model with multiple suppliers, just like water, electricity and gas.
Bruce Kornfeld
A8 Edge computing will surpass cloud for market share. Edge will be the prominent location that data and applications are managed and processed
James Maguire
@brucekornfeld To "edge eats the cloud" turns out to be right?
Mark Thiele he/him
A8 We won't know it's there, but it will be everywhere. Our assumption will be that things we need (to see, interact with, discover, use, etc) will just always and everywhere, be instantly available, with the flawless interaction you would expect from a desktop app.
BMC Software
A8: Focus will shift away from focusing on 'Device Edge' heavy use cases towards a more balanced set of use cases and delivering better experiences. (1/3)
Bruce Kornfeld
I'd say so. Cloud doesn't go away. but will have its place. Organizations need real-time processing and analytics....no matter how fast Cisco makes routers - you still have to wait for "speed of light" impact.
BMC Software
A8: In 10 years+ we won’t view #EdgeComputing as a separate ‘edge case’ (pun intended). Instead, Edge Computing will be a ubiquitous and commoditized element of the core fabric of computing. (2/3)
Lakshmi Sharma
No Edge does not eat the Cloud. Edge is THE Cloud as well. Edge is for fast and Simple instant results, Edge CLoud and HyperScaler Clouds work in conjunction . IN between these two is a continuum of Zero trust Networks,/SASE Global / Data guidelines, and more.
BMC Software
A8: Massive growth in complexity will force a greater focus on common standards for interoperability and management. (3/3)
James Maguire
@ilakshmis I agree -- they will merge. One big omnipresent world.
Chris Ehrlich
A8: Edge computing will rival, if not overtake, cloud computing for corporate spend. Dynamic hardware will evolve to take advantage of the field junction and the tech innovation dependent on it.
Lakshmi Sharma
A8: Agree , some standardization on what what API data models, API Security practices, Data Governance policies, may exist., Integration and smooth migration of workloads across Edge (WASM like run time environment (Bytecode Alliance) ) with K8 environment will be solved.
James Maguire
Q5. So, Big Question #1 of 2: What do you see as the future of edge computing over the next 2-4 years?
Bruce Kornfeld
A5. The edge data problems will be addressed…..the industry needs innovation around storing, protecting and using data at the edge. We need to stop trying to "shoe horn" cloud/datacenter technologies into edge.
Charlie Ashton
A5: In the near term, edge computing deployments will be driven by use cases with actual demonstrable business value like video content delivery, retail, healthcare, agriculture, energy, manufacturing and broadband access.
Mark Thiele he/him
A5 It starts in mostly invisible solutions, you're using it, but don't know it. In 2 years there will be increased innovation at the client level, making it more visible. By 4 years differentiating app perfo/UX for customers will be the key driver for adoption
BMC Software
A5: In the near-term #EdgeComputing will continue to be primarily driven by ‘Edge In’ approaches, delivering specific solutions to meet targeted business objectives. This will be particularly true in ‘Device Edge’ heavy industries such as Automotive, Manufacturing&Utilities.(1/4)
BMC Software
A5: In the slightly longer term as #EdgeComputing continues to mature, we expect to see greater involvement of the hyperscalers, both delivering broad functionality via ‘Cloud-out’ approaches & supporting industry specific edge use cases through wider partnerships.(2/4)
BMC Software
A5: Most organizations will adopt a combination of ‘Cloud-out’ and ‘Edge-in’ specific solutions resulting in greater operational & technology complexity.(3/4)
BMC Software
A5: Growing complexity will necessitate further standardization & harmonization efforts to reduce the burden of integration, interoperability & complexity. (4/4)
Chris Ehrlich
A5: More edge data centers will be developed, telcos will roll out more edge connectivity solutions, and manufacturers with IoT devices will incorporate edge into their planning.
Lakshmi Sharma
A5: Growth with personalization, SEO optimization, Authorization Offloads, Cost Optimization etc. and technically we see move towards, declarative and intent driven orchestration to expedite innovation through rapid prototyping
James Maguire
Q3. What are the most challenging edge computing issues that are shaping its future?
Bruce Kornfeld
A3.Protecting data at the edge is difficult - ……
Using data at the edge is difficult…..most tools are written for datacenters or cloud

(edited)

Mark Thiele he/him
A3 From the beginning, one of the biggest issues to Edge deployments has been the economic model associated with deployment of small scale, widely distributed infrastructure.
BMC Software
A3: Management @ Scale: The vast number of deployed nodes, combined w/ the distributed nature of #EdgeComputing disrupts traditional approaches to management. Often deployed in hard-to-reach locations, or w/out access to local skills, successful operation demands autonomous (1/2)
BMC Software
A3: or remote management capabilities. (2/2)
Charlie Ashton
A3: Standard servers lack the compute power to handle edge workloads such as high-performance network packet processing, video analytics or AI inferencing within typical constraints around energy consumption and cost.
Lakshmi Sharma
A3: The market is fragmented. Definition for Edge is fragmented. Folks coming from 5G / Telco see the world of Edge differently then the folks from Hyperscalers. And then there are companies like Fastly who have been focused on App and Developers ,
Bruce Kornfeld
@mthiele10 agreed. and the cloud folks are trying to build this for customers - it's just too expensive
James Maguire
@mthiele10 Are you saying small scale deployment is more challenging?
BMC Software
A3: (1/2) Data Management & #Analytics: The ephemeral nature of data combined w/a low signal to noise ratio necessitates being able to dynamically assess junk vs. value at scale. This is further compounded by the growing presence of #AI at the Edge in support of Edge Analytics.
Bruce Kornfeld
@BMCSoftware these sites often don't even have reliable internet connections.....there needs to be a new paradigm for management for sure
Chris Ehrlich
A3: Reliable high-speed connectivity for real-time data processing, remote device and end-user security, and edge hardware for compute.
BMC Software
A3: (2/2) An emerging challenge is managing the lifecycle of #AI across #edgecomputing devices.
Lakshmi Sharma
Some of the fellow panelists mention that it is the compute power needed for hosting or for large scale data cebters is not what you need at Edge. You need tech. e.g. serverless for scaling and WASM, for handling web level requests and security you need at Web
Bruce Kornfeld
@cdashton and there typically isn't a lot of space available - this is driving some of the innovation with GPUs, DPUs, etc..
Mark Thiele he/him
@brucekornfeld Frankly, I don't think the #Cloud folks will be the solvers of this problem, at least not in the near term.
James Maguire
Q2. What key trends are driving edge computing here in 2022?
Charlie Ashton
A2: Cloud edge infrastructure is increasingly being deployed not just in traditional on-premise data centers, but in new metro edge data centers, in edge devices themselves (e.g. cameras, vehicles, planes) and even integrated into telecom infrastructure (towers).
Bruce Kornfeld
A2. The need to process and analyze edge data where it is created to improve customer experience and improve efficiency/profits. Cost, latency and reliability issues with cloud are also driving edge thinking.
BMC Software
A2: Analytics @ The Edge: Traditional centralized analytics approaches do not scale effectively (or accommodate edge latency and bandwidth demands). When combined with the need to make decisions faster, this pushes analytics as close to the source of the data as possible.
Charlie Ashton
A2: And users are increasingly able to manage and consume services on edge infrastructure the same way as they do on traditional cloud infrastructure, with orchestration tools delivering application elements when and where they’re needed.
Lakshmi Sharma
A2: Personalization. Edge is important for fast innovation to personalize the offerings an enterprise or a customer has. Innovate so you can bring services faster to the market , and innovate so you can support the instant growth that you are looking for..
Chris Ehrlich
A2: The maturing front-line things, or devices, in the Internet of Things (IoT), untapped real-time big data analytics for business outcomes, and storage demand.
Bruce Kornfeld
@BMCSoftware completely agree! this is a huge area of focus for many users and vendors out there
Lakshmi Sharma
@brucekornfeld @bmcsoftware agree, for making decisions fast for your users going to the internet now and watching that game now or doing that financial transactions now.. you need data close to the app and the user. Data at edge offers the best outcome for your user

(edited)

James Maguire
Q7. What’s a big myth associated with edge computing and its future?
Mark Thiele he/him
A7 That the large hyperscalers will end up owning it.
BMC Software
A7: There is only one form of #edgecomputing: it’s either in the core or the edge.
In reality #Edge computing is a complex and diverse set of technologies stretching across a continuum from #cloud to edge.
James Maguire
@mthiele10 Who will "own" it if not the leading hyperscalers?
Charlie Ashton
A7: “Edge” is not the same as “on-premise”. “Edge” implies cloud connectivity + services, though that connectivity may be intermittent in many applications. Cloud provides scalability, virtualization, on-demand provisioning, automation and consumption-based pricing.
Bruce Kornfeld
A7 Myth: "The edge doesn't apply to me". Edge computing is everywhere - retail stores, manufacturing sites, healthcare (clinic and hospitals). It's simply the concept that processing and data is happening outside the datacenter and cloud
Bruce Kornfeld
@cdashton And think about it - all those things you say about cloud - they don't exist at the edge. I feel bad for the IT teams out there - its like the wild, wild west out at the edge with limited tools available!
Mark Thiele he/him
I don't know that I can speak to specific companies, but I do believe that innovation from smaller more agile players will determine much of what the Edge becomes. Some of the hardest questions to answer; Standards and Developer buy in.
Charlie Ashton
@brucekornfeld Yeah, big challenges and big business opportunities.
Chris Ehrlich
A7: Viewing the edge independent of cloud and data center infrastructures.
Lakshmi Sharma
A7: That it is only for IOT devices or it is for 5G.Then there are people who are moving up the stack from Infra to App.. they believe that Edge it to put Service Mesh or Service Discovery Tooling on Platform. While the Edge really is bringing these worlds together
Bruce Kornfeld
@mthiele10 right! big myth. They are trying.....but I don't see how they can innovate in these small sites.
James Maguire
@brucekornfeld They are motivated to be "the big players" and they have the deep pockets to try.
Lakshmi Sharma
@mthiele10 agree.. Edge is driven by innovation and depending on which industry you are in, their needs may be different. Some use cases we see consistently are Personalize Manifest manipulation for my content based on Geo, Compliance routing, Fast User Experience