Download to read offline
WR
Uploaded byWesley Reisz
189 views
Cloud to Edge
The document discusses the transition to edge computing and the implications of a rapidly growing data stream from vehicles, predicting 10 exabytes per month by 2025. It outlines the architecture of edge computing, comparing traditional and cloud-native approaches, and emphasizes the importance of adopting flexible, DevOps-driven strategies that leverage insights at both the device and infrastructure edges. Key takeaways highlight the necessity for developers to lead in maximizing the potential of edge computing, addressing challenges while applying cloud-native practices.
Related topics:
Recommended
![Vibe Coding vs. Spec-Driven Development [Free Meetup]](/image.pl?url=https%3a%2f%2fcdn.slidesharecdn.com%2fss_thumbnails%2fvibecodingvsspecdrivendevelopment-251209105622-43f455e7-thumbnail.jpg%3fwidth%3d640%26amp%3bheight%3d640%26amp%3bfit%3dbounds&f=jpg&w=240)
Cloud to Edge
- 1.Cloud to Edge:Architectingfor theNext Generation of theInternetWESLEY REISZVP Technology at SectionChairperson QCon SF &Co-host of The InfoQ Podcast
- 2.Toyota predicts thatthe car-to-cloud data stream will reach 10 exabytes per month by 2025.~2 million vehicles sold in North America alone
- 3.WESLEY REISZVP Technology@Section.ioChairperson QCon SF &Co-host of The InfoQ Podcastwes@sec&on.io@wesreisz
- 4.Section is adeveloper-focused EdgeCompute Platform for deployingcontainerized workloads at the edge.
- 5.DEFINE EDGEEdge overloadedtermARCHITECTURELet’s compare what things look likein and outside of an Edge Native ArchitectureCLOUD NATIVE LESSONSAT THE EDGEDevOps, ELK, K8s, Containers
- 6.Now is theMme to adopt an edge na'vestrategy at the device and infrastructureedge that is flexible, leverages DevOps,and offers deep insights into theopera7on.TL;DR
- 7.
- 10.LASTMILEDEVICEEDGEINFRASTRUCTUREEDGESelf-containedend-point devices, suchassmart-phones, wearables andautomobiles.Gateway devices such as IoTaggregators, switching and routingdevices.On-premise server platforms.Access sites which housenetwork access equipment,such as cellular radio basestations, xDSL and xPON accesssites.Aggregation hubs, such as those which house DAS (DistributedAntenna Systems) and serve as an initial aggregation oftransmission connections from the access sites.Regional data centers and central offices, where access controller,switching equipment and other service gateway functionality
- 11.CENTRALIZEDCLOUDINFRASTRUCTUREEDGEDEVICEEDGEIOT/SENSORSHigh Latency /More ComputeLower Latency / Less Compute
- 13.A content delivery network or contentdistribution network (CDN) is ageographically distributed network of proxy servers and their data centers.The goal is to provide high availability and high performance by distributingthe service spatially relative to end-users…CDNs are a layer in the Internet ecosystem. Content owners such as mediacompanies and e-commerce vendors pay CDN operators to deliver theircontent to their end users. In turn, a CDN pays ISPs, carriers, and networkoperators for hosting its servers in their data centers.Edge Compute Platformcompute, and custom workloads,^Edge Compute Platforms Application Developersensure workloadsare running between their servers and users.An edge compute platform leverages available compute between the cloudprovider and the user (together with cloud devops practices) to deploytraditional CDN & custom workloads.
- 14.Use CasesFacebook Live:Streamis sent via RTMPS (Real-Time Messaging Protocol) to a geographically local PoP. Theconnection is forwarded over an internal Facebook network to a Facebook data-centre. Whenyou see a live stream in your feed and you click on it the player requests the manifest. If it isn'talready on your local PoP the request goes to the data centre to get the manifest, and thenfetches the media files in 1 sec clips. As they get sent back they are cached on the PoP if theyaren’t there already.https://www.infoq.com/podcasts/sachin-kulkarni-facebook-live/Cloudflare:Service Worker API implementation for the Cloudflare platform. Brings a server less styleapproach to running JavaScript workloads on their Points of Presence. https://www.infoq.com/presentations/cloudflare-workers/Chick fil-A:We think of our Edge Computing environment as a “micro private cloud”. By this, we mean thatwe provide developers with a series of helpful services and a place to deploy their applicationson our infrastructure.https://medium.com/@cfatechblog/edge-computing-at-chick-fil-a-7d67242675e2Tesla:A Virtual Power Plant (VPP) is a network of distributed energy-resources (often solar, wind, andbatteries) that are aggregated to provide smarter and more flexible power generation,distribution, and availability. Tesla's VPP consists of vertically integrated hardware andsoftware, including both cloud and edge computing.https://qconlondon.com/london2020/presentation/tesla-virtual-power-plantInfrastructure Edge Compute
- 15.
- 16.
- 21.
- 22.
- 23.
- 24.CHALLENGESShared infraLimited Choice(often one size fits all)Often Little insight into what’s happening
- 25.
- 26.Points of Presences(POPs or Clusters)•OpEx vs CapEx•Dynamic•Multi-provider (all cloud vendorsDigitalOcean, RackCorp, CenturyLink, Packet)•Built on Top of the same technologies andpractices as the cloud•Containerized
- 27.Modules•Unit of deploymentis the container•Can horizontally scale workloads•Configuration is simplified•Choice of vendors without locking•Your own code can be deployWAFBOTContainer
- 28.CHALLENGESOrchestration of thesecontainersApplying the DevOps PracticesInsight into what’s happeningConnectivity
- 29.LESSONS FROM THECLOUDATTHE INFRA EDGE
- 30.FlexibilityConfigurable, ubiquitous,multi cloud/multi providerDevOps PracticesContinuous delivery/automation/containerized workloadInsights and VisibilityUnderstanding what’s happeningwith your system Observability
- 31.Developers need theability to package anddeploy workloads wherever they need logic toreside. They also need the tools to abstract thecomplexity of deploying to multiple edgelocations.We call this Any Workload, Any WhereFlexibility
- 32.RUNNING CONTAINERS… isa standard unit of software that packages up code and allits dependencies so the application runs quickly and reliablyfrom one computing environment to anotherProvides the isolation / dependencies /securityStandardizedLightweight
- 33.KUBERNETESKubernetes Namespaces allowsIsolation… is an open-source container-orchestration system forautomating application deployment, scaling, and management.It was originally designed by Google, and is now maintained bythe Cloud Native Computing Foundation.Manages the horizontal scale / availabilityComplicated, but understood at the cloud
- 34.MULTIPROVIDER DEPLOYMENTSTerraform, bash,API/CLINew languages/Tools (DarkLang andPulumi)Infrastructure as code (IaC) is the process of managing andprovisioning computer data centers through machine-readabledefinition files, rather than physical hardware configuration orinteractive configuration tools. The IT infrastructure managedby this comprises both physical equipment such as bare-metalservers as well as virtual machines and associatedconfiguration resources.
- 35.
- 36.
- 37.DevOps is the combinationof cultural philosophies,practices, and tools that increases an organization’s ability todeliver applications and services at high velocity: evolvingand improving products at a faster pace than organizationsusing traditional software development and infrastructuremanagement processes.This speed enables organizations to better serve theircustomers and compete more effectively in the market.You Build, You Own ItDevOps Practices
- 38.API DRIVEN… ableto write code to control aspects of the system.CustomizableOpenIntegration
- 39.GITOPS… is away of implementing Continuous Deployment for cloudnative applications. It focuses on a developer-centricexperience when operating infrastructure, by using toolsdevelopers are already familiar with, including Git andContinuous Deployment tools.GithooksTrigger message deployments to globalPOPSConfiguration management
- 41.
- 42.Things will eventuallygo wrong. When they do,developers need the ability to dive into the stack andunderstand what is happening. It should be familiarand combine the power of search and visualization.Insights and Visibility
- 43.OBSERVABILITYIn control theory,observability is a measure of how wellinternal states of a system can be inferred from knowledge ofits external outputs. The observability and controllability of asystem are mathematical duals3 Pillars of Observability :• Logs• Metrics• Tracing
- 44.
- 49.Connectivity… we canonly go so far today. K8s needsconnectivity. There are projects (such as Project EVEK3s, KubeEdge) that are pushing orchestration intothe device edge.Challenges Remaining
- 50.DEFINE EDGEEdge overloadedtermARCHITECTURELet’s compare what things look likein and outside of an Edge Native ArchitectureDevOps, ELK, K8s, ContainersCLOUD NATIVE LESSONSAT THE EDGE
- 51.KEY TAKEAWAYS•Edge isdivided into two groups defined by the last mile. On oneside is the device edge and the other is the Infrastructure edge.•Traditional applications have used CDN as a blackbox but morecan be done by applying some of the lessons learned at theCloud.•Challenges exist in leveraging cloud native practices at theInfrastructure edge; however, it is possible to leverage DevOps toaddress many of these issues.
- 53.An Industry Ripefor DisruptionRealizing the full potential of the Edge needs a catalyst- software developers -
- 54.QUESTIONSWESLEY REISZVP Technologyat SectionChairperson QCon SF &Co-host of The InfoQ Podcastwes@sec&on.io@wesreisz
- 55.
- 56.ServerlessContainerMODULESWeb App Firewall(WAF)Signal SciencesThreatXTemplarbitShieldImage OptimizationKrakenCloudinaryCacheVarnish 3-6Bot BlockerPerimeterX Bot BlockerImperva/ Distil BotBlockerCodeNode.jsOpenResty (Lua/nginx)Bring Your OwnA/B TestingSiteSpectContainerized AppsServerless Apps