SlideShare a Scribd company logo
1 of 28
extending cloud computing to the edge
Michel Burger
CTO
A personal journey
CSF
Service Syndication
Web Technology Pioneer
Web Service Interface on top of telecommunication equipment
Web Services and Pervasive Systems
A Path towards pervasive computing: A network approach
Device Broker (Device as a service)
User as a service
Distinguished Engineer
Vodaplex, GSMA One Identity,MEC
TMF Distinguished Fellow
Micro Gateway, Equinix Cloud Exchange
Edge Cloud Engine back end support
Software is evolving everything
1970-80: Assembler and developing compilers vs. high level language(s)
1980-90: Databases vs. software package (structured data)
1990-2000: Developing from scratch vs. using app server/middleware (ESB)
2000-2010: Scalable infrastructure vs. IT cloud resources (ala Amazon)
2010-2020: Big Data solution vs. a cloud solution; app service/middleware
vs.APIs and PaaS (ala Google appengine, force.com…)
2020-…: Cluster management vs Kubernetes/Istio; deployment infrastructure
vs deployment as a service; microservice (cloud edge) vs client/server
Evolution of core developers versus the others
Subroutines: make the code more readable
Pitfall: exiting the subroutines via GOTO…..
Local Libraries: benefit from the work done by others on the same environment
Pitfall: circular reference on import
Remote Libraries (Object): benefit from the work of others
Pitfall: hiding distribution (CORBA…)
Web Services: internal accessible Services
Pitfall: too complex
APIs: consuming a resource without having to worry on how it is implemented
Pitfall: clear understanding of differences between implementation and
exposure
Evolution of software structures
Evolution of the computing model
OS
Application
Resources
Disk-Centric Computing
Disk OS’s
DBMS
Spreadsheet
Word Processor
Device Drivers
(disk, tape, serial)
From 1950’s
Net-Centric Computing
TCP / IP
Telnet
FTP
HTTP
SMTP
Network Drivers
(3Com, Cisco,...)
From 1980’s
API-Centric Computing
API exposure
Mobile App
Multi Screen App
Client Mashup
Cloud Mashup
Service
Implementing API
From 2000’s
Agent-centric Computing
Analytics based
Agents
Fluid logic
Non-Imperative
Logistics
Devices/
Basic Services
(implemented as
agents)
Future
Disruption: Pressure on IT
IT
Digital
Assets
CMO asks for:
Mobile Apps, Partner Onboading
CFO requests:
“Lean”, Focus on core competencies
API
Data
Digital service ecosystem
where
services are internal or
external
Horizontalization where
every companies become a
service provider
Hybrid Cloud where
service provider is irrelevant
Everybody wants:
Consumerization on IT services
Build for Failure: Think someone yanking a VPC without warning – while you are working on it. When you start it back up again – it
should continue working without a hitch, no loss of customer data. DON’T STORE ANY STATE ON THE LOCAL MACHINE!
Build for Infinite Scale: Always assume there will be multiple instances of your code running on many machines.
Build for Latency: Network is slow, services are slow. Build your components not to be tied to state (I.E. Don’t hold state in memory
or local disk – unless you can recover it after a failure) whilst you are waiting and also use APM down to the last IO port (in most of our
cases a call into the network).
Build for Brewer Conjecture: You can have at most 2 of Consistency, Availability & Tolerant Network Partitions. Cloud services
don’t demonstrate eventual consistency. Deal with it in code.
Build for the Cloud: when you use cloud services – there are no guarantees. You are not guaranteed that you will get all the data,
you are not guaranteed that you will get them in order, you are not guaranteed anything! Now build your app to work with this.
Build with clean Interfaces: Don’t build with RPC Hangover. Build clean hierarchical REST interfaces using the resource model.
Build Probabilistically: Do we absolutely need to be a 100% accurate? Really? Are you sure? Positive? Use Bloom filters and caches
where possible.
Build to manage: Apply the FCAPS model to your services, other services/platform will need it.
Build to Cache: think HTTP caching and multiple layers of it. All GET operations must be cacheable.
Build to be Idempotent: same operation twice should not put your system in an inconsistent state.
Build for geo: build your component knowing that you could have multiple geographical partitions of data that you might have to use
sitting in different parts of the world.
Build for Data: make sure that what you write generates monetizable data.
Cloud architecture principles
Load based
Trigger: load changes (up or down)
Action: solution makes a request to an explicit broker for new IT resources or to release IT resources (request criteria is location and
resource type)
Resilience
Trigger: IT resource not available
Action: solution makes a request to an explicit broker for new IT resources as a replacement
Expansion of trigger and request criteria
Trigger: going beyond load and availability (eg: heat)
Action: solution makes a request to an explicit broker for new IT resources or to release IT resources (request criteria not only based on
location and resource type : eg heat).
Expansion to the implicit broker
Trigger: same as level 3
Action: the response is not limited to a specific data center and the resource returned can be within a different data center
Analytic based triggers and analytic based implicit requests
Trigger: analytics defines when to trigger the solution
Action : the solution makes a request to an implicit broker with a wide range of criteria and the IT resource returned is based on analytics
criteria
Evolution of cloud elasticity
Disruption: everything is getting easier
To Design
To Deploy
To Operate
API and languages (node.js)
Virtualization and Containers
Consumerization(API, Hybrid cloud)
and Elasticity
Disruption: everything is getting smarter
Data
From the consumers
From the services
From IT
From the network
Value
To Design
To Deploy
To Operate
API and languages (node.js)
Virtualization and Containers
Consumerization(API, Hybrid cloud)
and Elasticity
Microservice blueprint
Microservice
API
Database
Stream
Logs
Info
Publish/Subscribe or Queue
Events-out
Events-in
Other
API
(internal
or
external)
API Calls using
System tokens
Load Balancer
API Gateway
Cluster
management
agent
Sidecar
To expose API
to 3rd parties
Cluster
manager
Discovery
API
to
other
instances
in
the
cluster
API Calls using system tokens
As a library or external process to
handle the microservice as part of
a cluster of microservices
As a library or external process
to perform routing, control,
monitoring and micro API
gateway operations
Immutable infrastructure principles
Automation: Automate the setup and deployment for every part and
every layer of the infrastructure
No Partial Update: Never change any part of the system once it is
deployed. If it needs to be changed, deploy a new system.
Rolling Upgrade: Deployments will be done as rolling upgrades, and
therefore:
All database schema changes must be backwards compatible at least up to 2
versions
No Direct Access: Instances are headless with no SSH or direct access.
Automate sending all debug logs, etc. to centralized services
Microservice do and don’t
API First API is not SQL, generating a nice swagger file from code
does not really help
All APIs are secure via token No public/private APIs, avoid API gateway monolith
Configuration via environment variables or API No manual deployment or human access to production
machines
Microteam organization No technology transfer
Stick to one stack No rushing to the new shiny objects (Cargo cult syndrome),
polyglot systems are difficult to maintain
Test early Over testing and relying on test for quality
Logs must tell a story Breakpoints
Qualify your data (logs, info, events) No share of info between microservices except via API or
events, avoid data monolith
Horizontal scalability (Custer management) Scalability is not an option
System view (example)
1
Real Time
Communication
Contextual
Information
Charging &
Billing
Identity &
Authentication
Data & Info
Mgmt
Network
Optimization
Send Message Device Status Charge-to-bill Single Sign On Device management Network Load
Receive Message Device Capabilities Charge the provider Anonymous Customer
Reference
Provisioning
- Internal
- External
Best time of day
Call Control Location via Terminal Tariff Information Generic Validation Provisioning
Notification
- Internal
- External
Pre-load content
to CDN
Call Notification Device Connection Charge Notification Support of OpenID Address Book Push Notification
Update Call
- Add party
- Add medium
Location via Network Billing History Support of OAuth Backup
Transcoding
(Media Rendering)
Device Status
Notification
Payment
(Credit Card Processing)
Authentication for
Wifi Access
Resource Discovery
Transcoding
Notification
Location Notification
(Zone Based)
Payment
Configuration
Product Catalogue
- Internal
- External
User Profile Order Handling
User Profile
Notification
Order Handling
Notification
Device Connection
Notification
Product Catalogue
Notification
M-Data
(analytics, advertisement)
Sim Management
χ y
x
API Exposed via the Vodafone Platform
α N
β N
β RA
MN
β A
β A
α A
MRA
α A
β A
MRA
MRA
α A
M A
M
β A
α
M
α
1
2(1) 3 1 2
2 1 2 2(1)
2 2(1) 3 2
2 1 2 2
2(1) 1
Currently Available
Available by end of FY 12/13
Future Goals
3+
3+
MAN
MAN
MN
Mondrian
Update
NFC
Location Enabler
RCS-e
RCS-e
RCS-e
Smart Network
Smart Network
Messaging
Messaging
Mondrian
Mondrian
Mondrian
Location Enabler Mondrian
Mondrian
Mondrian
Mondrian
Mondrian
Mondrian Mondrian
Mondrian
Mondrian
Mondrian
Mondrian
Mondrian
Smart Network
Smart Network
M
Smart Network
Mondrian ER
ER Identity Mgmt
Identity Mgmt
Identity Mgmt
Identity Mgmt
Identity Mgmt
Identity Mgmt
Mondrian
Smart Charging
M-Analytics
Update
Content Proposition
Content Proposition
α N
1
2(1)
βRA
2
3+
3+
3+
3+
3
1
2
Beta5 project wave
Level of Interoperability:
- Redirection
- Application
- Network
Exposure Phase
β A
Parlay X (90%)
GSMA ONE API
Modelled on Parlay
X, SID, eTOM
Infrastructure abstraction: to go beyond virtualization to provide service with a set of APIs on how to consume basic IT
resources (Storage, Computing, Queuing, Network)
Federated Identity: based on open standard (OpenID connect, OauthV2), to deal with identity and authorization of
Subscribers, Employee and Partner for different legacy systems and handle identity from 3rd party service
Federated Profile: to act as a discovery of all data resources for specific top level entities (like users) as entry points to
distributed graphs of data
Consistent Operation Management (Distributed Transaction Management): to provide the infrastructure, top level APIs
and API specifications for system components to deal with distributed transactions for consistent operations like user
or service lifecycle
Eventually Consistent Operation Management (High Performance Notification/Event Management): to handle
eventually consistent real-time, near real-time distributed notification/event system for services using the notion of
inbox, outbox queues
Analytics: to handle via a multi-structured stream, all data extracted from every services/system components (as source
of data) and provide thru API is a way to consume the data from the stream
Exposing service APIs in the proper context: services are about API and not just end to end experience
Pillars of a microservice based platform
Federated Identity
Becoming an Identity Provider -> OpenId Connect)
• Consent Management
Federating internal identities -> (OauthV2)
Single Sign-On
• Entry point of a web graph via Top Level Entity
• Indicates the location of user’s disparate data resources
• Combines multiple resources into one view of data across multiple systems
Federated profile
…
Gold mining is about dirt management
0.05 ounces/ton
Analytics analogy I
About 11850 Amps to
generate around 8.4 Tesla
fields (about 150000 times the
earth magnetic field) but they
operate at low Voltage
A lot of what LHC is about is electricity flow management
Analytics Analogy II
Analytics is like LHC combined with gold extraction
• Huge amount of data -> 6.6 Zbytes/year by 2016 (Cisco Cloud Index)
• Big flow of data -> 400TB/day (Facebook)
• LHC generates 10-15 Petabytes/year of data for each experiment
Analytics
SSD memory (Violin), Redis
S3, Glacier
Snowflake, Athena, Redshift
Kinesis, Kafka
• Buffering, Routing, Filtering
• Structured/Unstructured store
• Event Collector
• Batch Process/Multi Structure Stream
• Multi Stage Store/Process
• Memory, disk, tape, archive based Non
Real-time
Real-time
Low
level
Semantic
Rich
Semantic
Cheap
Storage
High
Efficient
Storage
Analytic architecture
API
Data card
Sim Card
Content
Services
Knowledge
References
Applications
Social Networks
Connected Things
(Consumer,
Enterprise)
Connected Devices
Network Core
IT Infrastructure
RAN
Premise
Gateway
• Many different data structures
• Many different ways to extract the data
• Many different locations (even for the same type of data)
• Batch and Realtime based
• Buffered or stream
• Correlation parameters
Report Statistics
Data as a Service
Graph Network
Analysis
Neural Network
Analysis
Sources
Stream
Machine Learning
• Structured Buffering
- Proprietary
- Graph
- Neural Network
- Relational
• Unstructured Buffering
- Streaming
• Taping at Source
• Taping on Stream
• Consumption to
Source
Cloud analogy
Collocation Centralization of production
Decentralization of consumption
Mass distribution of production
and consumption
Collocation is back
Dam used as cheap storage
Evolution of cloud: edge cloud
‣ edge cloud enables new device types to act as
servers
‣ routers, gateways, mobile devices, game consoles
‣ serverless microservices may run on any device
‣ edge cloud is local
‣ closer to the user, closer to the data
‣ peer to peer communication
‣ edge cloud is multi-server
‣ apps decide where their “backend” services run
‣ central cloud, local machine
‣ edge cloud is distributed and dynamic
‣ edge servers may be ephemeral, mobile
‣ edge cloud scales
‣ new nodes add more resources at the edge
Disruption: from client/server to application/service
The Web normalized the client-server approach in designing and deploying solutions:
Microservices (cloud or edge) exposing API and deployed on demand at the right place are
normalizing the ability to create solutions by composing services:
Server
Client
device network cloud
application µ
service
µ
service
µ
service
µ
service
µ
service µ
service
device network cloud
Disruption: from star to hybrid topology
application
µ
service
µ
service
device
application
µ
service
µ
service
application
µ
service
µ
service
µ
service
network
µ
service
µ
service µ
service
cloud
Devices are not attached to the cloud, but
for an extension of the cloud at the edge
New types of clusters are becoming
important to manage:
• network
• proximity
• account
Edge cloud principles
• Meritocracy: all nodes can participate but take different roles based on
their merit
• Decentralised discovery: nodes discover each other based on scopes:
network, proximity & account
• Clustering: based on scopes network, proximity & account
• Microservice communication: nodes can communicate directly at
microservice level
• Dynamic resource instantiation: resources are made available only as
needed
• Independence: nodes can communicate regardless of operating
system, network, location
central cloud edge cloud
an edgeEngine & edgeBackend to enable devices to act as cloud servers.
a pragmatic approach to edge cloud
edgeEngine
• downloadable SDK for most
operating systems.
• devices can act as cloud
severs to other devices.
• devices can form clusters and
share resources.
edgeBackend
• backend for global
management, registration,
authorization, authentication
and security.

More Related Content

Similar to Cloud to hybrid edge cloud evolution Jun112020.pptx

Cloud Computing Networks
Cloud Computing NetworksCloud Computing Networks
Cloud Computing Networksjayapal385
 
Understanding the Windows Azure Platform - Dec 2010
Understanding the Windows Azure Platform - Dec 2010Understanding the Windows Azure Platform - Dec 2010
Understanding the Windows Azure Platform - Dec 2010DavidGristwood
 
Istio Service Mesh
Istio Service MeshIstio Service Mesh
Istio Service MeshLew Tucker
 
Fanug - Pragmatic Windows Phone Developer
Fanug - Pragmatic Windows Phone DeveloperFanug - Pragmatic Windows Phone Developer
Fanug - Pragmatic Windows Phone DeveloperSam Basu
 
Sampling from the Cloud Smorgasbord
Sampling from the Cloud SmorgasbordSampling from the Cloud Smorgasbord
Sampling from the Cloud SmorgasbordJim O'Neil
 
Cloud Computing Final1
Cloud Computing Final1Cloud Computing Final1
Cloud Computing Final1Sandip Kadam
 
Real-time Visibility at Scale with Sumo Logic
Real-time Visibility at Scale with Sumo LogicReal-time Visibility at Scale with Sumo Logic
Real-time Visibility at Scale with Sumo LogicAmazon Web Services
 
SaaS Enablement of your existing application (Cloud Slam 2010)
SaaS Enablement of your existing application (Cloud Slam 2010)SaaS Enablement of your existing application (Cloud Slam 2010)
SaaS Enablement of your existing application (Cloud Slam 2010)Nati Shalom
 
Azure Overview Arc
Azure Overview ArcAzure Overview Arc
Azure Overview Arcrajramab
 
Sukumar Nayak-Agile-DevOps-Cloud Management
Sukumar Nayak-Agile-DevOps-Cloud ManagementSukumar Nayak-Agile-DevOps-Cloud Management
Sukumar Nayak-Agile-DevOps-Cloud ManagementSukumar Nayak
 
Technology Overview
Technology OverviewTechnology Overview
Technology OverviewLiran Zelkha
 
Harbour IT & VMware - vForum 2010 Wrap
Harbour IT & VMware - vForum 2010 WrapHarbour IT & VMware - vForum 2010 Wrap
Harbour IT & VMware - vForum 2010 WrapHarbourIT
 
WP7 & Azure
WP7 & AzureWP7 & Azure
WP7 & AzureSam Basu
 
Monitoring as an entry point for collaboration
Monitoring as an entry point for collaborationMonitoring as an entry point for collaboration
Monitoring as an entry point for collaborationJulien Pivotto
 
Cloud Computing & Sun Vision 03262009
Cloud Computing & Sun Vision 03262009Cloud Computing & Sun Vision 03262009
Cloud Computing & Sun Vision 03262009guest829442
 
Role of cloud and analytics in IoT
Role of cloud and analytics in IoTRole of cloud and analytics in IoT
Role of cloud and analytics in IoTSelvaraj Kesavan
 

Similar to Cloud to hybrid edge cloud evolution Jun112020.pptx (20)

Cloud Computing Networks
Cloud Computing NetworksCloud Computing Networks
Cloud Computing Networks
 
Understanding the Windows Azure Platform - Dec 2010
Understanding the Windows Azure Platform - Dec 2010Understanding the Windows Azure Platform - Dec 2010
Understanding the Windows Azure Platform - Dec 2010
 
Istio Service Mesh
Istio Service MeshIstio Service Mesh
Istio Service Mesh
 
Fanug - Pragmatic Windows Phone Developer
Fanug - Pragmatic Windows Phone DeveloperFanug - Pragmatic Windows Phone Developer
Fanug - Pragmatic Windows Phone Developer
 
Sampling from the Cloud Smorgasbord
Sampling from the Cloud SmorgasbordSampling from the Cloud Smorgasbord
Sampling from the Cloud Smorgasbord
 
Cloud Computing Final1
Cloud Computing Final1Cloud Computing Final1
Cloud Computing Final1
 
Real-time Visibility at Scale with Sumo Logic
Real-time Visibility at Scale with Sumo LogicReal-time Visibility at Scale with Sumo Logic
Real-time Visibility at Scale with Sumo Logic
 
Intro Cloud Computing
Intro Cloud ComputingIntro Cloud Computing
Intro Cloud Computing
 
SaaS Enablement of your existing application (Cloud Slam 2010)
SaaS Enablement of your existing application (Cloud Slam 2010)SaaS Enablement of your existing application (Cloud Slam 2010)
SaaS Enablement of your existing application (Cloud Slam 2010)
 
Azure Overview Arc
Azure Overview ArcAzure Overview Arc
Azure Overview Arc
 
Sukumar Nayak-Agile-DevOps-Cloud Management
Sukumar Nayak-Agile-DevOps-Cloud ManagementSukumar Nayak-Agile-DevOps-Cloud Management
Sukumar Nayak-Agile-DevOps-Cloud Management
 
Technology Overview
Technology OverviewTechnology Overview
Technology Overview
 
Harbour IT & VMware - vForum 2010 Wrap
Harbour IT & VMware - vForum 2010 WrapHarbour IT & VMware - vForum 2010 Wrap
Harbour IT & VMware - vForum 2010 Wrap
 
WP7 & Azure
WP7 & AzureWP7 & Azure
WP7 & Azure
 
Monitoring as an entry point for collaboration
Monitoring as an entry point for collaborationMonitoring as an entry point for collaboration
Monitoring as an entry point for collaboration
 
Cloud Computing & Sun Vision 03262009
Cloud Computing & Sun Vision 03262009Cloud Computing & Sun Vision 03262009
Cloud Computing & Sun Vision 03262009
 
Role of cloud and analytics in IoT
Role of cloud and analytics in IoTRole of cloud and analytics in IoT
Role of cloud and analytics in IoT
 
Azure Microsoft
Azure MicrosoftAzure Microsoft
Azure Microsoft
 
htcia-5-2015
htcia-5-2015htcia-5-2015
htcia-5-2015
 
Cloud Computing
Cloud ComputingCloud Computing
Cloud Computing
 

More from Michel Burger

New Design Patterns in Microservice Solutions
New Design Patterns in Microservice SolutionsNew Design Patterns in Microservice Solutions
New Design Patterns in Microservice SolutionsMichel Burger
 
Service syndication
Service syndication Service syndication
Service syndication Michel Burger
 
Light reading conference
Light reading conferenceLight reading conference
Light reading conferenceMichel Burger
 
The path towards pervasive computing
The path towards pervasive computingThe path towards pervasive computing
The path towards pervasive computingMichel Burger
 

More from Michel Burger (8)

New Design Patterns in Microservice Solutions
New Design Patterns in Microservice SolutionsNew Design Patterns in Microservice Solutions
New Design Patterns in Microservice Solutions
 
Embrace network
Embrace networkEmbrace network
Embrace network
 
Active network
Active networkActive network
Active network
 
Service syndication
Service syndication Service syndication
Service syndication
 
Light reading conference
Light reading conferenceLight reading conference
Light reading conference
 
The path towards pervasive computing
The path towards pervasive computingThe path towards pervasive computing
The path towards pervasive computing
 
Sdp summit
Sdp summitSdp summit
Sdp summit
 
Cloud mobility
Cloud mobilityCloud mobility
Cloud mobility
 

Recently uploaded

Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationSlibray Presentation
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr LapshynFwdays
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupFlorian Wilhelm
 
costume and set research powerpoint presentation
costume and set research powerpoint presentationcostume and set research powerpoint presentation
costume and set research powerpoint presentationphoebematthew05
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesSinan KOZAK
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxnull - The Open Security Community
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationSafe Software
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks..."LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...Fwdays
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Scott Keck-Warren
 
Unlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power SystemsUnlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power SystemsPrecisely
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Mattias Andersson
 
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersEnhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersThousandEyes
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphNeo4j
 

Recently uploaded (20)

Vulnerability_Management_GRC_by Sohang Sengupta.pptx
Vulnerability_Management_GRC_by Sohang Sengupta.pptxVulnerability_Management_GRC_by Sohang Sengupta.pptx
Vulnerability_Management_GRC_by Sohang Sengupta.pptx
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck Presentation
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project Setup
 
costume and set research powerpoint presentation
costume and set research powerpoint presentationcostume and set research powerpoint presentation
costume and set research powerpoint presentation
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen Frames
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
 
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort ServiceHot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
DMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special EditionDMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special Edition
 
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks..."LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024
 
Unlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power SystemsUnlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power Systems
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?
 
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersEnhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
 

Cloud to hybrid edge cloud evolution Jun112020.pptx

  • 1. extending cloud computing to the edge Michel Burger CTO
  • 2. A personal journey CSF Service Syndication Web Technology Pioneer Web Service Interface on top of telecommunication equipment Web Services and Pervasive Systems A Path towards pervasive computing: A network approach Device Broker (Device as a service) User as a service Distinguished Engineer Vodaplex, GSMA One Identity,MEC TMF Distinguished Fellow Micro Gateway, Equinix Cloud Exchange Edge Cloud Engine back end support
  • 3. Software is evolving everything
  • 4. 1970-80: Assembler and developing compilers vs. high level language(s) 1980-90: Databases vs. software package (structured data) 1990-2000: Developing from scratch vs. using app server/middleware (ESB) 2000-2010: Scalable infrastructure vs. IT cloud resources (ala Amazon) 2010-2020: Big Data solution vs. a cloud solution; app service/middleware vs.APIs and PaaS (ala Google appengine, force.com…) 2020-…: Cluster management vs Kubernetes/Istio; deployment infrastructure vs deployment as a service; microservice (cloud edge) vs client/server Evolution of core developers versus the others
  • 5. Subroutines: make the code more readable Pitfall: exiting the subroutines via GOTO….. Local Libraries: benefit from the work done by others on the same environment Pitfall: circular reference on import Remote Libraries (Object): benefit from the work of others Pitfall: hiding distribution (CORBA…) Web Services: internal accessible Services Pitfall: too complex APIs: consuming a resource without having to worry on how it is implemented Pitfall: clear understanding of differences between implementation and exposure Evolution of software structures
  • 6. Evolution of the computing model OS Application Resources Disk-Centric Computing Disk OS’s DBMS Spreadsheet Word Processor Device Drivers (disk, tape, serial) From 1950’s Net-Centric Computing TCP / IP Telnet FTP HTTP SMTP Network Drivers (3Com, Cisco,...) From 1980’s API-Centric Computing API exposure Mobile App Multi Screen App Client Mashup Cloud Mashup Service Implementing API From 2000’s Agent-centric Computing Analytics based Agents Fluid logic Non-Imperative Logistics Devices/ Basic Services (implemented as agents) Future
  • 7. Disruption: Pressure on IT IT Digital Assets CMO asks for: Mobile Apps, Partner Onboading CFO requests: “Lean”, Focus on core competencies API Data Digital service ecosystem where services are internal or external Horizontalization where every companies become a service provider Hybrid Cloud where service provider is irrelevant Everybody wants: Consumerization on IT services
  • 8. Build for Failure: Think someone yanking a VPC without warning – while you are working on it. When you start it back up again – it should continue working without a hitch, no loss of customer data. DON’T STORE ANY STATE ON THE LOCAL MACHINE! Build for Infinite Scale: Always assume there will be multiple instances of your code running on many machines. Build for Latency: Network is slow, services are slow. Build your components not to be tied to state (I.E. Don’t hold state in memory or local disk – unless you can recover it after a failure) whilst you are waiting and also use APM down to the last IO port (in most of our cases a call into the network). Build for Brewer Conjecture: You can have at most 2 of Consistency, Availability & Tolerant Network Partitions. Cloud services don’t demonstrate eventual consistency. Deal with it in code. Build for the Cloud: when you use cloud services – there are no guarantees. You are not guaranteed that you will get all the data, you are not guaranteed that you will get them in order, you are not guaranteed anything! Now build your app to work with this. Build with clean Interfaces: Don’t build with RPC Hangover. Build clean hierarchical REST interfaces using the resource model. Build Probabilistically: Do we absolutely need to be a 100% accurate? Really? Are you sure? Positive? Use Bloom filters and caches where possible. Build to manage: Apply the FCAPS model to your services, other services/platform will need it. Build to Cache: think HTTP caching and multiple layers of it. All GET operations must be cacheable. Build to be Idempotent: same operation twice should not put your system in an inconsistent state. Build for geo: build your component knowing that you could have multiple geographical partitions of data that you might have to use sitting in different parts of the world. Build for Data: make sure that what you write generates monetizable data. Cloud architecture principles
  • 9. Load based Trigger: load changes (up or down) Action: solution makes a request to an explicit broker for new IT resources or to release IT resources (request criteria is location and resource type) Resilience Trigger: IT resource not available Action: solution makes a request to an explicit broker for new IT resources as a replacement Expansion of trigger and request criteria Trigger: going beyond load and availability (eg: heat) Action: solution makes a request to an explicit broker for new IT resources or to release IT resources (request criteria not only based on location and resource type : eg heat). Expansion to the implicit broker Trigger: same as level 3 Action: the response is not limited to a specific data center and the resource returned can be within a different data center Analytic based triggers and analytic based implicit requests Trigger: analytics defines when to trigger the solution Action : the solution makes a request to an implicit broker with a wide range of criteria and the IT resource returned is based on analytics criteria Evolution of cloud elasticity
  • 10. Disruption: everything is getting easier To Design To Deploy To Operate API and languages (node.js) Virtualization and Containers Consumerization(API, Hybrid cloud) and Elasticity
  • 11. Disruption: everything is getting smarter Data From the consumers From the services From IT From the network Value To Design To Deploy To Operate API and languages (node.js) Virtualization and Containers Consumerization(API, Hybrid cloud) and Elasticity
  • 12. Microservice blueprint Microservice API Database Stream Logs Info Publish/Subscribe or Queue Events-out Events-in Other API (internal or external) API Calls using System tokens Load Balancer API Gateway Cluster management agent Sidecar To expose API to 3rd parties Cluster manager Discovery API to other instances in the cluster API Calls using system tokens As a library or external process to handle the microservice as part of a cluster of microservices As a library or external process to perform routing, control, monitoring and micro API gateway operations
  • 13. Immutable infrastructure principles Automation: Automate the setup and deployment for every part and every layer of the infrastructure No Partial Update: Never change any part of the system once it is deployed. If it needs to be changed, deploy a new system. Rolling Upgrade: Deployments will be done as rolling upgrades, and therefore: All database schema changes must be backwards compatible at least up to 2 versions No Direct Access: Instances are headless with no SSH or direct access. Automate sending all debug logs, etc. to centralized services
  • 14. Microservice do and don’t API First API is not SQL, generating a nice swagger file from code does not really help All APIs are secure via token No public/private APIs, avoid API gateway monolith Configuration via environment variables or API No manual deployment or human access to production machines Microteam organization No technology transfer Stick to one stack No rushing to the new shiny objects (Cargo cult syndrome), polyglot systems are difficult to maintain Test early Over testing and relying on test for quality Logs must tell a story Breakpoints Qualify your data (logs, info, events) No share of info between microservices except via API or events, avoid data monolith Horizontal scalability (Custer management) Scalability is not an option
  • 15. System view (example) 1 Real Time Communication Contextual Information Charging & Billing Identity & Authentication Data & Info Mgmt Network Optimization Send Message Device Status Charge-to-bill Single Sign On Device management Network Load Receive Message Device Capabilities Charge the provider Anonymous Customer Reference Provisioning - Internal - External Best time of day Call Control Location via Terminal Tariff Information Generic Validation Provisioning Notification - Internal - External Pre-load content to CDN Call Notification Device Connection Charge Notification Support of OpenID Address Book Push Notification Update Call - Add party - Add medium Location via Network Billing History Support of OAuth Backup Transcoding (Media Rendering) Device Status Notification Payment (Credit Card Processing) Authentication for Wifi Access Resource Discovery Transcoding Notification Location Notification (Zone Based) Payment Configuration Product Catalogue - Internal - External User Profile Order Handling User Profile Notification Order Handling Notification Device Connection Notification Product Catalogue Notification M-Data (analytics, advertisement) Sim Management χ y x API Exposed via the Vodafone Platform α N β N β RA MN β A β A α A MRA α A β A MRA MRA α A M A M β A α M α 1 2(1) 3 1 2 2 1 2 2(1) 2 2(1) 3 2 2 1 2 2 2(1) 1 Currently Available Available by end of FY 12/13 Future Goals 3+ 3+ MAN MAN MN Mondrian Update NFC Location Enabler RCS-e RCS-e RCS-e Smart Network Smart Network Messaging Messaging Mondrian Mondrian Mondrian Location Enabler Mondrian Mondrian Mondrian Mondrian Mondrian Mondrian Mondrian Mondrian Mondrian Mondrian Mondrian Mondrian Smart Network Smart Network M Smart Network Mondrian ER ER Identity Mgmt Identity Mgmt Identity Mgmt Identity Mgmt Identity Mgmt Identity Mgmt Mondrian Smart Charging M-Analytics Update Content Proposition Content Proposition α N 1 2(1) βRA 2 3+ 3+ 3+ 3+ 3 1 2 Beta5 project wave Level of Interoperability: - Redirection - Application - Network Exposure Phase β A Parlay X (90%) GSMA ONE API Modelled on Parlay X, SID, eTOM
  • 16. Infrastructure abstraction: to go beyond virtualization to provide service with a set of APIs on how to consume basic IT resources (Storage, Computing, Queuing, Network) Federated Identity: based on open standard (OpenID connect, OauthV2), to deal with identity and authorization of Subscribers, Employee and Partner for different legacy systems and handle identity from 3rd party service Federated Profile: to act as a discovery of all data resources for specific top level entities (like users) as entry points to distributed graphs of data Consistent Operation Management (Distributed Transaction Management): to provide the infrastructure, top level APIs and API specifications for system components to deal with distributed transactions for consistent operations like user or service lifecycle Eventually Consistent Operation Management (High Performance Notification/Event Management): to handle eventually consistent real-time, near real-time distributed notification/event system for services using the notion of inbox, outbox queues Analytics: to handle via a multi-structured stream, all data extracted from every services/system components (as source of data) and provide thru API is a way to consume the data from the stream Exposing service APIs in the proper context: services are about API and not just end to end experience Pillars of a microservice based platform
  • 17. Federated Identity Becoming an Identity Provider -> OpenId Connect) • Consent Management Federating internal identities -> (OauthV2) Single Sign-On
  • 18. • Entry point of a web graph via Top Level Entity • Indicates the location of user’s disparate data resources • Combines multiple resources into one view of data across multiple systems Federated profile …
  • 19. Gold mining is about dirt management 0.05 ounces/ton Analytics analogy I
  • 20. About 11850 Amps to generate around 8.4 Tesla fields (about 150000 times the earth magnetic field) but they operate at low Voltage A lot of what LHC is about is electricity flow management Analytics Analogy II
  • 21. Analytics is like LHC combined with gold extraction • Huge amount of data -> 6.6 Zbytes/year by 2016 (Cisco Cloud Index) • Big flow of data -> 400TB/day (Facebook) • LHC generates 10-15 Petabytes/year of data for each experiment Analytics
  • 22. SSD memory (Violin), Redis S3, Glacier Snowflake, Athena, Redshift Kinesis, Kafka • Buffering, Routing, Filtering • Structured/Unstructured store • Event Collector • Batch Process/Multi Structure Stream • Multi Stage Store/Process • Memory, disk, tape, archive based Non Real-time Real-time Low level Semantic Rich Semantic Cheap Storage High Efficient Storage Analytic architecture API Data card Sim Card Content Services Knowledge References Applications Social Networks Connected Things (Consumer, Enterprise) Connected Devices Network Core IT Infrastructure RAN Premise Gateway • Many different data structures • Many different ways to extract the data • Many different locations (even for the same type of data) • Batch and Realtime based • Buffered or stream • Correlation parameters Report Statistics Data as a Service Graph Network Analysis Neural Network Analysis Sources Stream Machine Learning • Structured Buffering - Proprietary - Graph - Neural Network - Relational • Unstructured Buffering - Streaming • Taping at Source • Taping on Stream • Consumption to Source
  • 23. Cloud analogy Collocation Centralization of production Decentralization of consumption Mass distribution of production and consumption Collocation is back Dam used as cheap storage
  • 24. Evolution of cloud: edge cloud ‣ edge cloud enables new device types to act as servers ‣ routers, gateways, mobile devices, game consoles ‣ serverless microservices may run on any device ‣ edge cloud is local ‣ closer to the user, closer to the data ‣ peer to peer communication ‣ edge cloud is multi-server ‣ apps decide where their “backend” services run ‣ central cloud, local machine ‣ edge cloud is distributed and dynamic ‣ edge servers may be ephemeral, mobile ‣ edge cloud scales ‣ new nodes add more resources at the edge
  • 25. Disruption: from client/server to application/service The Web normalized the client-server approach in designing and deploying solutions: Microservices (cloud or edge) exposing API and deployed on demand at the right place are normalizing the ability to create solutions by composing services: Server Client device network cloud application µ service µ service µ service µ service µ service µ service device network cloud
  • 26. Disruption: from star to hybrid topology application µ service µ service device application µ service µ service application µ service µ service µ service network µ service µ service µ service cloud Devices are not attached to the cloud, but for an extension of the cloud at the edge New types of clusters are becoming important to manage: • network • proximity • account
  • 27. Edge cloud principles • Meritocracy: all nodes can participate but take different roles based on their merit • Decentralised discovery: nodes discover each other based on scopes: network, proximity & account • Clustering: based on scopes network, proximity & account • Microservice communication: nodes can communicate directly at microservice level • Dynamic resource instantiation: resources are made available only as needed • Independence: nodes can communicate regardless of operating system, network, location
  • 28. central cloud edge cloud an edgeEngine & edgeBackend to enable devices to act as cloud servers. a pragmatic approach to edge cloud edgeEngine • downloadable SDK for most operating systems. • devices can act as cloud severs to other devices. • devices can form clusters and share resources. edgeBackend • backend for global management, registration, authorization, authentication and security.

Editor's Notes

  1. Today the API gateway and microservices are hosted on the central cloud. So even for an application to run in an standalone mode, it has to send and receive requests through central servers.
  2. How it works … We provide a SDK (called edgeSDK) that is available on GitHub, and is downloadable to most computing devices and is available for most operating systems. Once the developer downloads the SDK, it created a cloud environment on the device for the developer to develop microservices and application on the top. It enables the application to communicate directly across devices, and enables devices to form clusters for sharing resources. And we have a powerful backend that provides the global functions such as managing the devices, authentication and security.