Part 0: StresS Tests And Scaling Woes

What is a Stress Test

• Intentionally placing higher than average load on an entity. This helps to understand the performance of whatever is being stress tested under heavy load conditions.

What is Scaling

Scaling is a process for increasing the capacity of a system to address growing needs and resource constraints and generally has two flavors

• Horizontal Scaling - Add more servers
• Vertical Scaling - Add more resources

The importance Of Failure

• Experience
• Resiliency
• Improvement

PostMORTEM Review

• Blameless culture
• Purpose
• Structure
  • Date
  • RCA
  • Trigger
  • Resolution
  • Detection
  • AI (Action Items)
  • LL
    • What went well
    • What went poorly
    • Where we got lucky
  • Timeline

Part1: Service-Oriented Architectures

History and SOA introduction

What Is a Service?

Real world 'service':

• A piece of work performed by a service provider

• Takes some input and produces some desired results

  • e.g. a restaurant: pay some money and get some food

  • e.g. a roofing company: pay some money, get a roof repair

• Has quality characteristics (price, time, goodness of product, etc.)

Software world 'service':

• Takes some input, performs some work, produces some output
• Request-response model: client requests, server responds
• Has quality characteristics (price, execution time, availability, security,  goodness of product, etc.)

Definition

A "web service" is a piece of software that performs processing and uses a web protocol to accept requests and issue responses.

earlier...

web service: a piece of software that serves up data through a web interface. Typically web services are object-oriented, provide access to a database, and encode data in XML or JSON.

Definition

SOA (Service-Oriented-Architecture) is an architectural paradigm that modularizes business functions into services by decoupling and encapsulating different portions of the business logic into different service components.

• Services in SOA:
  • are autonomous and stateless
  • accept requests and return responses
  • use well-defined, standard interfaces (standard protocols) that define inputs and output structures
  • platform independent
  • discoverable

Autonomous -> Need not be aware of other services

stateless -> need not remember state from request to request

• Improves scalability through horizontal cloning
• Can store state in a database if it needs to be saved

standard interfaces -> Re-use web stack and define custom application endpoints

credit:

Torsten Braun, Universität Bern

platform independent -> Write a service once, support many platforms

​

discoverability

​

Service Actors

Services need consumers. Consumers use a client to make requests to the service's server. They expect to get a desired result or an error message.

Service providers host and manage their service offerings - maintaining service quality and security.

Service brokers match service consumers and providers - sometimes composing multiple services to fulfill consumer needs.

Modularization and Service orientedness

Modularization Through Service-orientedNess

Will come back to this

img credit: http://usblogs.pwc.com/emerging-technology/agile-coding-in-enterprise-it-code-small-and-local/

Service Types

1. Infrastructure as a service (IaaS)
2. Platform as a service (PaaS)
3. Software as a service (SaaS)
4. Desktop as a service (DaaS)
5. Service as a service (SaaS2)

Example

credit: https://m.oursky.com/saas-paas-and-iaas-explained-in-one-graphic-d56c3e6f4606

Part2: Service Oriented Methods/protocols 
and Data formats

A Tale of Two Eras

Legacy Services:

• Uses SOAP, WSDL, XML, WS-*, etc

• Older, more formally specified service types

Modern Services:

• RESTful API services:
  • Uses HTTP, REST, JSON
  • Most pervasive services circa 2020
• "Streaming" services:
  • Uses websockets
  • real time applications with "server push"
    usecases
• New age services:
  • ​graphQL
  • gRPC (google Remote procedure call)

SOAP ERA

RESTful ERA

Elements of a "Legacy" Service

Description in terms of WSDL (Web Service Definition Language):

• Describes what a web service can do
• WSDL is an XML based, open standard from W3C
• Declares available methods (endpoints)
• Identifies Input and output parameters

Protocol uses SOAP to exchange XML, XSD on top of HTTP

• SOAP (Simple Object Access Protocol)

• XML (Extensible Markup Language)

• XSD (Extensible Schema Definition)

HTTP

Example WSDL

SOAP

Simple Object Access Protocol

• Encapsulated in an Envelope
• SOAP Header
• SOAP Body
• operates over HTTP or TCP
• Allows for requests and responses
• fill in parameters according to WSDL

SOAP Request Example

Note: Refers to some function called 'CalcDistance' that has two parameters 'startPoint' and 'endpoint'. Each point has an x and y field in the object.

SOAP ReSponse Example

Modern Services

REST

• Representational State Transfer
• Associate every resource with a URI (Universal Resource identifier) - accessible from a URL
• Allow CRUD operations to retrieve and persist data -> works really well with SOA
• We've seen REST at work with services like Twitter
• REST has lots of advantages of most other approaches, but can result in overfetching or underfetching

img credit: https://www.howtographql.com/basics/1-graphql-is-the-better-rest/

graphQL

• Basic idea: Traverse a graph to get only the data you need
• Solves the problem of underfetching or overfetching
• Invented by facebook
• downside:
  • complicates architectural design (harder to use micro services)
  • very difficult to cache, since all requests are unique

img credit: https://www.howtographql.com/basics/1-graphql-is-the-better-rest/

WebSockets

• Works directly on a single TCP connection
• Provides full duplexing (bi-directional) communication between server and client - remains open during the duration of the connection
• Great for streaming applications that rely upon a rapidly updating publish/subscribe model - particularly for server pushes 

DATA Formats

XML

• XML is a markup-language for data representation
• Used for encoding documents in machine-readable form
• Text-based format, consists of tags, attributes and content
• Can be used for data, meta-data, and structural presentation

JSON

• JavaScript Object Notation
• Basically a set of key:value pairs that define data according to a dictionary structure. 
• Typically much less verbose than XML
  • e.g. <somedatatype>somedatavalue</somedatatype> becomes { "somedatatype": somedatavalue}.
  • particularly true for nested structures
• Used for data structures not page or document structuring
• Can still include meta-data easily.

RPC/graphs/etc

• Remote procedure calls have been around for a long time. Depending on the RPC framework used, there are variant semantics. 
• Graphs are data representations of actual networks of connected objects. Data is often serialized into a JSON format, but may be held in a variety of different formats.