CDL-Slides Flashcards
Define Digital Transformation
The degree by which org leverage cloud technology to improve operations, workflows, and infrastructure to become more agile, efficient, and competitive in the digital age.
Define Cloud Computing
The practice of using a network of remote servers (computers), to store, manage, and process data versus a local or personal server (computer)
When hosting your infrastructure on-premis, what are you responsible for managing? What level of risk does this approach present?
You manage:
- real estate
- actual servers + maintenance and upgrades
-IT personnel
High level of security but you bear all risk
When a cloud provider manages your infrastructure, what are they generally responsible for? What level of risk does this approach present?
They assume responsibility of real estate, IT staffing, and servers.
Still secure, risk depends on SRM and type of cloud service
What is the evolution of computing to the cloud?
1) Dedicated Servers 2) Virtual Private servers 3) Shared hosting 4) Cloud Hosting
Define Cloud Hosting and its benefits.
Cloud hosting is a grouping/system of physical machines abstracted into cloud services.
Benefits: Cost effective, scalable, secure, flexible, and highly configurable.
Define a dedicated server. What are its limitations vs benefits?
One physical machine hosted on sight for a single business that runs a single web-app/site.
Pro: Highest security
Cons: Expensive, high maintenance
Define a Virtual Private Server. What are its limitations vs benefits?
One physical machine dedicated to a single business, but is broken down into sub-machines via virtualization.
Pro: Runs multiple web-app/sites
Cons:
Define Shared Hosting. What are its limitations vs benefits?
One machine shared by hundreds of businesses.
Pros: Very cheap
Cons: Very limited - banks on tenants under utilizing their resources.
What is a Cloud Service Provider?
A company that provides multiple cloud services, of which can be linked together to create cloud architectures.
Eg: Cloud CDN + VPC + CE + Cloud Storage + Cloud SQL, etc
What is Google’s first GCP service?
App Engine in 2008
What are the 7 benefits of cloud computing, regardless of provider? (Not on exam, but important)
- On-demand (PAYG) Pricing
- Global - launch workloads wherever in minutes.
- Security - Physical security manage by providers AND cloud services are secure by default AND ability to implement granular configuration/control
- Reliability - Back up, disaster recover, data replication, and fault tolerance.
- Scalability - Increase or decrease resources on-demand (e.g., machines, compute, or storage)
- Elastic - Automate scaling during spikes and drop on-demand (e.g., peek events)
- Current Hardware- Software always patched, upgraded, and replaced without interruption.
What 4 services contribute to reliability?
- Back up
- Data replication
- DIsaster Recovery
- Fault tolerance.
What are Google’s 4 major cloud computing services?
- Compute
- Networking
- Storage
- Databases.
What are the 3 types of cloud computing services on the Cloud Pyramid and who are their end users?
1 (Top) - SaaS / Clients, End users
2 (Middle) - PaaS / Developers
3 (Bottom) - IaaS / Admins
What benefit does PaaS offer developers?
They do not have to worry about
1. provisioning - obtaining necessary resources (servers, VM, cloud resources)
2. configuring - fine tuning and adopting those resources to work properly.
3. Understanding underlying hardware OR OS
What benefits do IaaS provide Admins?
They do not have to worry about:
1. Data center + Real estate
2. Staffing
3. Maintenance and hardware refresh
What is a Shared Responsibility Model?
A visualization that describes what cloud providers versus end users are responsible for.
Each provider has a specific SRM
Within the SRM who has the least responsibility? The most? Who is in the middle?
Least - SaaS
Most - IaaS
Middle - PaaS
Review image.
What is in the cloud versus of the cloud?
In the cloud - If you can configure or store it, then most likely you are responsible.
Of the cloud - If you cannot configure/store, then most likely you are not responsible (cloud provider is)
What are customers always responsible for in the SRM?
- Data
- Configuring access controls via permissions
Define FaaS
A serverless computing service that allows developers to build, run, and manage applications without worrying about underlying infrastructure.
Devs only pays for compute time
Within SRM, describe the different responsibilities of parties when using Bare Metal (Compute Engine) / IaaS
Client: OS Host Configuration + Hypervisor
Provider: Physical machine only
Within SRM, describe the different responsibilities between parties when using Virtual Machine (Compute Engine) / IaaS
Client: Guest OS Configuration, Container Runtime,
CSP - hardware, storage, networking.
Within SRM, describe the different responsibilities between parties when using GKE/Containers / IaaS
Client: Configuration, deployment, and storage of containers.
Google: OS, Hypervisor (docker daemon?), Container Runtime
Within SRM, describe the different responsibilities between parties when using Managed Platforms, like App Engine / IaaS
Client: Code, deployment strategies, environment configuration, and associated services
Google: Servers, OS, Networking, Storage
Within SRM, describe the different responsibilities between parties when using SaaS?
Client: Content, files, access controls
Google: Servers, OS, Networking, Storage, Security
Within SRM, describe the different responsibilities between parties when using FaaS, such as Cloud Functions?
Client: Uploading code
Google: Basically everything, deployment, container runtime, networking, storage, security, server
What 4 cloud services are needed to run SaaS?
Compute, storage, database, cloud networking (VPC, subnets, etc.)
Can VM be both sole and multi-tenant?
Yes
Sole-tenant - When you are the only person using that specific server, machine
Multi-Tenant - when multiple tenants are using host machine/server
What are the 3 different Cloud Computing Deployment models?
- Public Cloud (cloud native)
- Private Cloud (on-prem)
- Hybrid Cloud (on-prem + native)
Define Public Cloud (aka cloud native).
Use to describe when everything is built on the cloud
Define Private Cloud
AKA, On-premise, when everything is built on company datacenters.
Define Hybrid Cloud. Whats a use case?
The use of on-prem and cloud providers (e.g., moving data to the cloud)
Define Cross-Cloud
AKA, multi-cloud, when you are using and running workloads across multiple cloud providers
E.g: GCP Anthos
What types of clients tend to use Cloud?
Start ups
What types of clients tend to use Hybrid?
Banks, FinTech, Professional services, legacy on-prem users.
Why? BC data protections and privacy due to nature of data
What type of clients tend to use on-premise?
PubSec, Gov, Hospitals, Large enterprises with heavy regulations (insurance)
What are the different costs associated with CAPEX (on-prem) versis OPEX (GCP)?
CAPEX: Physical costs & software licenses
OPEX: Subscription costs and operational billing metrics (compute, storage, networking, etc.)
Define and explain difference costs between versus CAPEX and OPEX?
CAPEX (software license fees)- Upfront spending on physical infrastructure, such as servers, storage (hard drives), network (routers, cables, switches), backup/archive, disaster recovery, datacenter (rent, security, cooling), and IT staff
OPEX (subscription fees) - Non-physical costs related to operations, such as Software leasing and customizations/features, cloud support, training staff on cloud services, and cloud metrics (e.g., compute and storage usage)
What are the benefits/limitations between CAPEX and OPEX?
CAPEX - Deduct expenses from tax bill. CON: You have to guess upfront what you plan to spend.
OPEX: You can try a product/service without investing in equipment.
Define Availability. Provide an example.
The ability of your service to remain available (operational) by ensuring there is no single point of failure and/or ensuring a certain level or performance.
Ex: Cloud load balancing - The distribution of traffic to multiple servers in one or more data center. Running your workloads across multiple zones ensure that if 1 or 2 zones fail, the load balancer will rout traffic to the datacenter still operational.
Define Scalability, and explain the difference between vertical and horizontal scaling.
Your ability to increase your capacity based on increased demand or traffic, memory or computing power.
Vertical (scaling up) - Upgrade to bigger server.
Horizontal (scaling out) - Adding more of the same server of the same size, and then distributing workloads across those machines.
Define Elasticity
The ability to increase/decrease capacity based on current demand of traffic, memory, and computing power (happens automatically versus scalability)
NOTE: Vertical scaling is harder for traditional architectures, so you’ll generally only see horizontal scaling described with elasticity.
Define fault tolerance. Provide an example.
The ability for your service to ensure this is no single point of failure, focused on preventing chance of failure.
Eg: Having a secondary, redundant, database where all changes are synced automatically and remain on stand-by. Think B/G environment.
GCP Service: Cloud DNS - A service that detects a failing primary system and fail-over to a standy-by, secondary system.
Define horizontal scaling, and describe the difference between scaling OUT vs IN
The ability to increase/decrease capacity based on current demand of traffic, memory, and computing power.
Out - When you add more servers of the same size.
In - When you remove more servers of the same size.
NOTE:
How is elasticity achieved automatically with GCP?
Via Managed Instance Groups (MIGs) - a defined schedule that automatically increase/decrease capacity in response to a demand.
Define fail over. Which cloud concept is it associated with?
A operational plan to shift traffic to a redundant, secondary system in case the primary system fails.
E.g., Cloud DNS
Define High Durability
Your ability to RECOVER and PREVENT the loss of data from a disaster.
Define Disaster Recovery
Solutions that recover data from a disaster.
Do you have a back up? How fast can you restore that backup? Does your back up still work? How do you ensure current live data is not corrupt?