Migration to Cloud

  1. TCO or ROI ？

    • Total Cost of Ownership:即总拥有成本，包括产品采购到后期使用、维护的成本。这是一种公司经常采用的技术评价标准。
    • Return on investment: 投资回报率,是指通过投资而应返回的价值，即企业从一项投资活动中得到的经济回报。它涵盖了企业的获利目标。利润和投入经营所必备的财产相关，因为管理人员必须通过投资和现有财产获得利润。投资可分为实业投资和金融投资两大类，人们平常所说的金融投资主要是指证券投资

  2. 需要考虑的tech point

    • Security, bring in your secuirty specialists to audit your deployment
    • Compliance
    • Scale
    • Perforamnce
    • Resilency

  3. 需要的人员

    • cloud architects
    • cloud developers
    • cloud security engineers

  4. 需要改变的思想

    • Think Small
    • Think Simple
    • Budget for mistakes
    • Never be afraid to fail

  5. Cloud Security Planning:

    • deal with the basic: at least the minimum viable security
    • be proactive: able bo monitor, react and anticipate
    • provide clarity in security policies, standard, processes, roles and accountabilities

  6. CloudOps 是什么？

    • continues operations and continues improvement
    • zero downtime (zdt)
    • security monitoring is part of cloudOps

  7. Moving to the cloud means:

    • creating a scurity approach and technology solution that is appropriate for the applications and data stores

  8. Cloudops toolsets take into account:

    • the needs of the applications and data sets that are part of ops

  9. Operating security means:

    • being proactive

  10. useful links:

Basic to Mirosoft Azure

  1. Overview

  2. Resource Group

  3. Virtual Network

    • CIDR
    • 10.1.0.0/16 means 10.1.0.0 to 10.1.255.255, 65536 addresses
    • Subnet
    • 10.1.0.0/24 means 10.1.0.0 to10.1.0.255, 256 addresses
    • Microsoft suggest gateway subnet to /26 or /28
    • use peerings to connect to Azure Vnets
    • 5 ip address are used in an Azure subnet

  4. Storage

    • Container Blob service:
    • File service: smb3
    • the Azure storage exploer app must be download and install locally when bulk upload needed

  5. VM

    • On-premise
    • Cloud solution:
      • networking
      • storage
      • workloads
        • domain controller
        • DNS servers
        • line of business apps
        • remote desktop servies
      • high availability
      • backups
    • Hybrird implemention:
      • gateway subnet
      • vpn gateway
        • it is used to connect your vnet or your virtual network in Azure to your on-premise env
        • site to site
        • point to site: remote user
      • dedicated vpn appliance
      • expressroute: private connection between your network and Azure deployment

  6. Settings:

    • newtworking
    • storage, shape, scale
    • security
    • avavilablilty set or zone (HA)

  7. Ops:

    • backup
    • DR disaster recovery
    • monitor

  8. Azure active directory: cloud control panel

    • manage users, devices and apps through a single portal
    • backbone of IDM(indentity management in Azure)
    • users and group:
    • connect tool: connect on-premise env to Azure Active Directory

  9. Backup and DR:

    • DR:
    • RTO: recovery time objective, duration or lenghth of time that business mut be functional after disaster
    • RPO: recovery point bjective, maximum amount of data that can be lost after a disaster
    • Backup:
      • LRS: localy redundant storage, 3 copies of your data within the region
      • GRS: geo-redundant storage, data s replicated to another region

Basic to Google cloud platform(GCP)

  1. GCP services:

    • Compute
      • VM
      • Container, Docker
      • Functions and lambdas
    • Storage and DB
      • Files
      • Archive storage
      • Persistent disks for vms
      • data service
        • noSql
        • rdbms
        • hadoop and spark
    • Big dta services:
      • data pipelines
      • data warehouses
      • data science notbooks
      • machine learning
    • Other services:
      • identity and security:
      • manager and monitor
      • developer tools

  2. GCP 卖点

    • google is the speed demon
    • automatic scaling
    • dev tools and libs
    • 通过命令行创建和使用service而不是常看到的UI操作

  3. Steps to start GCP

    • setup account
    • navigate to google cloud console
    • work with localtions
    • billings
    • use APIs
    • IAM
    • gcloud, cli for gcloud
    • get demo code from git (gcp homwpage ->language)
    • 卖点 walk-throughs and tutorials，一步一步指导创建service
    • 卖点 gcp compare aws doc，可以查看Google云和亚马逊云的优劣点比较

  4. Computing services:

    • GCS google cloud storage
    • GCE google compute engine
      • vm instance, ssh is transfer by google, you do not need to keep it by chrome
      • using cloud launcher to setup an eclipse ide that means to have a browser-based IDE to code in java on a gcs vm
      • using gce resource
        • storage: image, snapshots, os versions
        • networking: networks, firewalls, routes
    • GKE google container engine
      • docker for app virtualization
      • kubernets for management : kubectl
    • GAE google app engine
      • for simeple westie, an old product

  5. Cloud storage:

    • storage JSON api (like swagger)
    • db, gql doc-style db
    • bigtable is designed to support wide column db, HBase, for logging senarios
    • bigquery data warehouse as a service
    • bigtable vs big query

  6. Data pipeline: commonly used for loading and storing data to analyze

    • cloud pub/sub messsaging: asyn, topic-based msg service
    • cloud dataproc: hadoop , spark…
    • cloud dataflow: apache beam, big ETL extract, transform and load
      • it’s a pipeline creation service that integrates with the other google cloud products such as
        • cloud storage for files
        • pub/sub for messages
        • datastore for docs
        • bigtable for log tables
        • bigquery for sql ourtput
      • it’s the glue between all the services
    • google genomics, used in pipelines with bigquery

  7. Machine learning and visualization

    • cloud vision api
    • cloud datalab: Ipython stype/ Jupyter stype for data science

  8. Netowrking and dev tools:

    • cloud networking service
      • network
      • external ip : to connect outside GCP
      • firewall rules
      • routes
      • lb
      • dnc
      • vpn
      • cloud routers
    • stackdriver monitoring: monitor your workloads
      • monitor GCP
      • monitor AWS
    • source code tools: enable private Git hosting on the GCP
    • what neat about this is they’re dockerized and so you can just simply run a docker comd and you can get a full configured env for dev

  9. Implementing solutions: how to combine the most commonly used Google Cloud services into buldabe arcitecture for our projects? 各种困惑中ing

    • reference artcitectures: cloud.google.com/icons
    • DR architecture
      • GCE
    • web/api app architecture
      • GCE and LB
      • GKE/GAE and container mangagement
      • Cloud DNS
      • Cloud SQL or other DB
      • [{< image classes=“figure nocaption fig-100” src=“/images/gcp/GCP-Archi-Web1.png” >}}
    • big data and data warehouse
      • cloud storage
      • bigquery
      • cloud dataflow or 3rd ETL
      • data studio or 3rd dashboards
    • internet of things
      • cloud storage
      • cloud pub/sub
      • cloud bigtable
      • cloud dataflow
      • bigquery
    • bioinformatics
      • cloud storage and public datasets
      • bigquery
      • GCE instance or cluster
      • dataflow
      • datalab notebooks
    • launch checklist for GCP berore you’re going to lanch your project commercially (deploy to production)