Cloud Elements

Cloud Elements

1. Cloud Infraestructure

   • Regions
     • EX: NA SOUTH US EAST
     • Availability Zones
       • *Having different zones decreases a single point of failure * Connected to other AZs and DataCenters

2. Computing Resources

   • Virtual Servers or Virtual Machines (VMs)
     • Software based servers that are virtualized.Also known as virtual instances or instances depending on the provider.
     • Public VMS
     • Transient or Spot VMs
       • * If theres unused capacity in this VM you can buy them at huge discounts but the provider can decommisionned them at any moment for higher price services * They are great for non production, testing and developing, among others.
     • Reserved virtual server instances
       • * Reserve capacity and guarantee resources for future deployments for a contracted term for a set number of years, * You will lower your costs compared to hourly or monthly rates. * You can commision public vms in case you exceed your capacity.
     • Dedicated Hosts
       • * They offer single tenant isolation, that means only your VMs run on a single host. * They are tipically used for compliance requirements or licensing terms.
     • Hybrid Multicloud
       • Refers to the deploy model where you use both private and public clouds from different providers, if they were from the same provider it would be a monocloud.
       • Cloud Scaling
         • Using a private cloud for a normal year performance but when theres peaks like in holidays, you can use public cloud to leverage their capacity and work well
         • Composite Cloud
           • Its a next step on the cloud scaling but taken for applications spread across multiple cloud enviroments.
           • Example: A company that has their app's (composed of UI, billing API and Rewards system) on premises server on Europe has been noticing that on american holidays its getting slow so they decide to compose their application across multiple cloud environments. Their solution was to move only their billing and UI to a cloud in an american datacenters so it can scale well enough in holidays, while their european premises individually scale.
           • https://learning.edx.org/course/course-v1:IBM+CC0101EN+2T2020/block-v1:IBM+CC0101EN+2T2020+type@sequential+block@73f2d32c3cc74ee79f26769dbe5f365b/block-v1:IBM+CC0101EN+2T2020+type@vertical+block@052a7b7f7e4c4c139411a61520c3ca38
       • Modernization
         • Example: Airlines use mobile apps with a cloud mobile backend + access to their in premises reservation server. If you could design a booking feature in the cloud for flight delays, users would be way happier.
         • Data + AI
           • Example: Since 30% of delay times in the airline industry is attributed to unplanned maintanence, you could use AI or machine learning in the public cloud to hook up to the legacy information in their private servers about unplanned maintanence to get predictive analytics about when errors or unplanned maintanence will occur.
   • Bare Metal Servers
     • Physical Servers that do are not virtualized. They are single tenant, dedicated physical servers.
     • * Fully Customizable * Dedicated or Long-term usage * High Performance Computing * Highly Secure / Isolated Enviroments. * No hypervisor * More expensive than VMs
   • Serverless
     • Offloads responsability for common infrastructure management tasks such as scaling, scheduling, patching, provisioning.
     • * No provisioning of servers and runtimes. * Runs code on-demand on a per request basis, scaling as needed. * PAY ONLY WHEN INVOKED AND USED.
     • IT DOESNT MEAN THERES NO SERVERS, IT MEANS THAT SERVER MANAGEMENT IS REMOVED FROM THE USERS.
     • Challenges: * Workloads characterized by long-running processes might be better with traditional server enviroments.
     • Use Cases
       • * Data and event processing * IOT * Microservices * Mobile Backends
       • Good Conditions for Serverless: * Short-running stateless functions. * Seasonal workloads * Production volumetric data that shows too much idle time. * Event-based processing. * Stateless microservices.

3. Storage

   • VM or BM both have local data storage buth if they are decommissioned the data is lost, so theres options for persistent data based on access speed, access frequency and security over the data.
   • File Storage
     • Also known as NFS (NETWORK FILE SYSTEM) Storage. It needs to be attached to a compute node via ethernet networks. It is recommended for tasks that do not depend on speed. Common workloads include repository of files or other tasks where speed variance is not and issue.
     • Disadvantages: * Slower Advantages: * Low Cost * Can be attached to multiple servers * Resilient to failute * Speeds vary based on network speed * More Secure
   • Object Storage
     • Is the most common way of storage in the cloud. Its great for backup and disaster recovery solutions and unstructured data, namely, data without any kind of hiererchical folder. They are stored in buckets tha cannot contain other buckets.
     • Not attached to a compute node, it uses an API instead. Ideal for data which is static and where fast read and write speeds are not necessary. NOT SUITABLE FOR OS, DATABASES OR CHANGING CONTENT.
     • Advantages: * Its infinite  (File and block are not) * Cheapest one Disadvantages: * Slowest one
     • Buckets
       • It has data tiers: * Standard: stores objects that are frequently accessed and has the highest cost * Vault / Archive: Stores objects that are accessed once or twice a month. * Cold Vault: Stores data that is only accesed once or twice a year, cheapeast one.
       • Automatic archiving rules for your data, namely, if some of you data is not being accessed for a long time its gonna go down in tiers so its cheaper to pay for its storage.
       • Does not come with IOPS options Cold vault bucket takes hours If you need fast access object storage is not the suitable one.
     • API
       • Does not need computing nodes, you access data via API. S3 API is the most common one. Normally providers use S3 or S3 compatible API so developers can access different providers object storage.
   • Block Storage
     • Advantages: * Faster than file storage * Extremely resilient to failure * Data is more secure * Higher price point than file storage * Workloads that need low latency and high speeds.
     • Must be connected to a node for it to work. Block storage breaks file into block that later will be storing separately into different addresses.
   • Direct Attached
     • * It needs to be attached to a compute node * Its fast * Normally only contains the OS * Its Ephemeral (Temporary) * Not Shared
   • IOPS
     • Input Output Operations per Second. Importan concept since if its too low you will be getting a bottleneck in your operations and if its too high you may not be using the full of it and paying more for it.

4. Networking

5. Virtualization

   • Creating a software based or "virtual" version of something
   • Hypervisor
     • Software that runs above the physical servers and manages and distributes the resources of these among your differents VMs
     • Type 1 / Bare Metal
       • *They are the most common *Have less latency than the type 2
     • Type 2 / Hosted
       • Frequently used for end user virtualization Higher Latency than type 1 They use host os between the phys. server and the hypervisor thats why they are called hosted
   • VMs
     • * They have operating systems and applications * Independent from one another * Can run multiple of them on a hypervisor. * You can run different os in the different VMs
   • Benefits
     • Cost Savings
       • You can run multiple VMs from one piece of infrastructure You dont have to maintain servers You dont run electricity costs
     • Agility and Speed
       • Since getting up a vm is easy and you can run test on in a lot simpler and quicker
     • Less Down Time
       • Since you can move your VMs across different hypervisors almost instantanously if one goes down you can move your VMs to another one

6. Content Delivery Networks (CDN)

   • Distributed Server Network that delivers temporary stored or cached of website content to users based on the users geographical location. Reduces the distance between your webiste visitors and your website servers.
   • Service that accelerates internet content delivery. Its main benefit is that makes your website faster. Its indirect benefit its the reduction of capacity that you would need in your main servers since you are creating caches all around the world making the users not having to overload your main server. Other benefits would be and increase in uptime and security through obscurity.

7. Microservices

   • Is a type of architecture where a single application is composed of many loosesly couple and independently deployable, smaller components or services.
   • * These services typically have their own stack running on their own containers. * They communicate with one another using APIs, Event Streaming and Message Brokers.
   • Benefits
     • * Application components can be developed independently by multiple developers. * Teams can use different stacks and runtime enviroments for different components. * Components facing too much load can be scaled independently instead of scale the whole application.

8. Containers

   • Executable unit of software in which application code is packaged, along with its libraries and dependencies in common ways so that it can be run anywhere, whether is on desktop, traditional IT or the cloud. 
   • THEY DO NOT NEED TO HAVE A GUEST OS UNLIKE THE VMs AND INSTEAD LEVERAGE ON THE FEATURES AND RESOURCES ON THE HOST OS.
   • * The process of containerization would be the manifest, the image and finally the creation of the container. * Also in your server you would use a runtime engine instead of a Hypervisor
   • Benefits
     • * Scalability: Since we de not need a guest os out application will need less capacity than a VM. * Portability: Since they can be run anywhere. * Faster dev to market speed.

9. Cloud Security

   • Cloud Security

10. Cloud Monitoring

    • Monitoring performance across an entire stack of apps and services can be time-consuming and draining on internal resources.
    • Cloud monitoring solutions asses data, apps and infrastructure for performance, resource allocation, network availability, compliance and security risks.
    • https://learning.edx.org/course/course-v1:IBM+CC0101EN+2T2020/block-v1:IBM+CC0101EN+2T2020+type@sequential+block@c61f97bfa0364c9d937cd527b074ac1e/block-v1:IBM+CC0101EN+2T2020+type@vertical+block@995b206e8a8b44d285d971a2563b6fa2
    • Includes
      • Strategies, practics and processes used for analyzing, tracking and managing services and apps.
    • Categories
      • Infrastructure
        • Help identify minor and large scale - failures so that developers can take corrective action
      • Databases
        • Help track processes, queries and availability of services to ensure accuracy and reliability
      • App performance and monitoring
        • Help improve user exp, meet app user SLAs to minimize downtime and lower operational costs.
    • Best Practices
      • Leverage end user exp. monitoring solutions. Move all aspects of infrastructure under one monitoring platform. Use monitoring tools that help track usage and cost Increase cloud monitoring automation Simulate outages and breach scenarios