1.
Lesson Keynote
Distributed Systems in Data Engineering
By: Oluwasegun Matthew | oadetimehin@terragonltd.com
Summary
1. Introduction to Distributed Systems
a. The concept of server-client architecture
b. Channel for Communication
c. Impact on Data Engineering at Scale
2. From Localhost to Production - things to watchout for...
3. Industry based Technologies/Tools in View
a. Messaging kits - RabbitMQ & Kafka
b. In Memory Data Caching - Redis & Aerospike
c. Data in Stream Tools - AWS Kinesis
d. Monitoring and Log Watch - CloudWatch
4. Summary - in class
5. Questions
Class Activity:​ ​Form 4 groups, choose from any of the messaging and in-memory data caching
tool, use this to ​create a resilient distributed system​ to fix the following problems:
- Crashing nature of e-Portal portal
- Exam records processing
Let’s Dive In...
1

2.
1. Introduction to Distributed Systems
According to Wikipedia through Google,
A distributed system in its most simplest definition is a group of computers working together as to
appear as a single computer to the end-user. These machines have a shared state, operate
concurrently and can fail independently without affecting the whole system’s uptime.
This is in line with ever-growing technological expansion of the world, distributed systems are
becoming more and more widespread. Take a look at the increasing number of available
computer technologies/innovation around, this is sporadically increasing, and this result in
intense computational requirement.
Yeah, Moore’s law proposed more computing power by fitting more transistors (which
approximately doubles every two years) into a simple chip using cost-efficient approach - cool,
but over the past 5 years, there has been little deviation from this - ability to scale horizontally
and not just vertically alone.
2

3.
The Concept of Server-Client Architecture
Client-server architecture(client/server) is a network architecture in which each computer or
process on the network is either a client or a server.
Just the way it is in a general world, activities is usually based on server/client relationship and
this isn’t different in technology too e.g Cashier/Customer, Bus Conductor/Passengers etc.
Another type of network architecture is known as a peer-to-peer architecture because each node
has equivalent responsibilities -​ but this isn’t what we are discussing today
The approach of breaking breaking larger application into chunks over a server-client
architecture can be explained with ​Microservices. ​Consider the cases below:
3

4.
Case 1 - Monolith: ​At the core of the application is the business logic, which is implemented by
modules that define services, domain objects, and events. Surrounding the core are adapters that
interface with the external world. Examples of adapters include database access components,
messaging components that produce and consume messages, and web components that either
expose API or implement a UI - this results in ​Monolithic Hell
4

5.
Case 2 - Microservices:​ Here we are tackling complexity, A service typically implements a set of
distinct features or functionality, such as order management, customer management etc. Each
microservice is a mini-application that has its own hexagonal architecture consisting of business
logic along with various adapters. Some microservices world expose an API that’s consumed by
other microservies or by the application’s client. Other microservices might implement a web UI.
At runtime, each instance is often a cloud VM or a Docker container.
5

6.
Quiz ​Give…
● Examples of a client/server relationship in real world
● Methods of binding two systems that you know
● Two architectures in which softwares are designed
● Major issue with Monolithic design
6

7.
Channel for Communication
When we have a decentralized system, it’s important for us to make these systems communicate
with one-another. The client/server architecture emphasis a producer/consumer computing
architecture where the server acts as the producer and the client as a consumer. The model of
communication can either be ​synchronous​ or ​asynchronous​. Each of this further broken into:
- API Mode
- Buffer Mode
API Mode: is a synchronous (or instant feedback) mode of communication. It usually used for
one-to-one type of communication through protocols like http, https, smtp, smpp etc.
Buffer Mode: is an asynchronous mode of communication, where feedback isn’t needed
immediately. It works for both one-to-one and broadcast communication. In this mode of
communication, a queuing/messaging/buffering system is placed in between these two systems
to manage flow of information. Here the following queuing algorithm is emphasized:
- FIFO (First In First Out)
- LIFO (Last In First Out)
- SJF (Shortest Job First)
- Round Robin
Impact on Data Engineering at Scale
Again, bringing the concept of distributed system into data Engineering...Hey, what’s data
engineering?
Data engineering is the act of building and managing information or “big data” infrastructure.
Data engineers create architecture that helps analyze and process data in the way it’s needed by
an organization, from data processing to creating a pipeline of data into lake and warehouse for
business value creation.
The following are some of the positive impacts of distributed system in data engineering:
- Creating resilient data architecture
- Easily managed systems
7

8.
- Security and control
- Reduced failure point
- Fault detection with ease
8

9.
Quiz ​Mention...
● 2 Queue algorithms you are familiar with
● Web Technologies that runs on HTTP protocol
9

10.
2. From Localhost to Production - things to watchout for..
When systems are built on development environment, a lot isn’t considered, this may be due to
experience, right information or un-envisaged circumstances. This implies that a perfect system
cannot be built at development stage until it’s tested in real-life scenario.
Sometimes, system overkill design might be a major flaw of the development phase, but the
production will really tell or not.
List of things to watch out:
- Unexpected spike in platform/technology usage - system overload
- Performance as a result of consistent platform usage
- Security of interconnected systems
- Extensibility of features
- Easy of deployment
10

11.
Enough of theoretical exposition, Let’s go practical…
3. Industry based Technologies/Tools in View
Here we shall talk about the different tools used in the industry to manage distributed system
Messaging Kits ​- e.g. RabbitMQ or Kafka
RabbitMQ is the most widely deployed open source message broker - ​https://www.rabbitmq.com/
Tutorial Guide (in PHP) - https://www.rabbitmq.com/tutorials/tutorial-three-php.html
11

12.
12

13.
13

14.
In Memory Data Caching ​- e.g. Redis or Aerospike
Redis is an open source in-memory data structure store used as a databse, cache and message
broker. It supports data structures such as strings, hashes, lists, sets, sorted sets with range
queries, bitmaps, hyperlogs etc.. - ​https://redis.io/
Documentation found here for PHP: https://github.com/amphp/redis
14

15.
Data in Steam Tools ​- AWS Kinesis
AWS Kinesis makes it easy to collect, process and analyze real-time streaming data so you can
get timely insights and react quickly to new information; owned by Amazon
- https://aws.amazon.com/kinesis/
15

16.
Monitoring and Logs Watch​ - CloudWatch
`
AWS Cloudwatch is a monitoring and management service built for developers, system
operators, site reliability engineers (SRE), and IT managers https://aws.amazon.com/cloudwatch/
Assessment
See class activity on the first page...
16

17.
Questions and Mentorship
For further questions, collaboration or mentorship, reach out:
Email: oadetimehin@terragonltd.com
Mobile: 07060514642
17