THE NEEDS FOR ARGOS

The Forth Industrial Revolution

As it is stated in in the meaning of the fourth industrial revolution following the flow of thinking, the keyword of forth industrial revolution (IR) is automation; the first IR was steam; the second IR was electricity; the third IR was communication.

Automation is known as the robotic manufacturing system, but now it needs more attention in the software industry.

The software industry is taking the initiative over hardware along with the transition from third industry to the fourth.

 

  • GE announces its transition to the software and analytics company in 2017.
  • IBM has already transformed itself from a hardware company to the software and service company from the 2000s.

 

Ironically, the software industry is less automated than hardware industry regardless of industrial transition.

In the process of a software development project, developers must intervene from every procedure to the test results, and involve in the release, deployment, and operation process.

Now we can infer the direction to win the competitiveness in the fourth IR, which is the automation of development and operation system.

Next table shows the transition of Industrial Revolution.

 

 

1784

187019692020
Industrial
Revolution
1st2nd3rd

4th

Energy

CoalPetroleumElectricityRenewable Energy
CommunicationTelegraphTelephoneInternet

IoT (Internet of Things)

Transportation

RailwayAutomobileSmartReplicator (Local Production)
InfluenceConsumer Goods and Light IndustryElectrochemistry, Heavy Chemical IndustryIT

Every Industry

Paradigm

MachineMass ProductionDigital

AI (Artificial Intelligence)

Value

codevoiceinformationData
 

Production
Mechanism

Factory system and mechanization
by steam engine
Mass production by standardization of work and division of laborAutomation revolution by industrial robots

Automation with AI
(Artificial Intelligence)
Low volume and multi-product production

 

This paradigm emphasizes that only companies will survive who meet the diverse needs of customers by shifting its traditional manufacturing methods to automated small volume production. So, every enterprise must meet digital transformation to change and make a revolution in their business model. In regards of software digital transformation include the change of architectures.

 

  • Today’s demands are simply not met by yesterday’s software architectures. from https://www.reactivemanifesto.org
  • Mobile generation requires different development methods and architectures.

 

Software companies have out-competed traditional firms and even acquire hardware companies.

You can compare the top five companies with the highest market cap in 2007 and 2017.

 

Rank

20072017

1

PetroChinaApple
2ExxonMobil

Google Alphabet

3

GEMS
4China Mobile

Amazon

5ICBC
(Industrial and Commercial Bank of China)

Facebook

 

Digital conversion is the way to ride on the changing paradigm, and companies are more likely to disappear without it.

DevOps

DevOps is one of the most critical factors in the software-based digital transformation.

DevOps is an enterprise software development methodology used to unify an agile relationship between Development and IT Operations.

 

Let’s take a look at the DevOps of the Unicorn Company to see how they change and improve the relationship by advocating better communication and collaboration between the two business units.

 

Unicorn Company and DevOps

 

Facebook

  • Continuous iterations and incremental development
  • Facebook developers work in independent and released small changes frequently.
  • Done is better than perfect.
  • Code wins arguments.
  • The best idea and implementation should always win. 

 

Google (Ten Things we know To Be True in Google)

  • Focus on the user and all else will follow
  • It’s best to do one thing really, really well.
  • Fast is better than slow.
  • Democracy on the web works.
  • You don’t need to be on your desk to need an answer.
  • You can make money without doing evil.
  • There is always more information out there.
  • The need for information crosses all borders.
  • You can be serious without a suit.
  • Great just isn’t good enough.

 

Netflix

  • High Availability, but move fast.
  • Tools encourage best practice, but freedom to do the right thing.

 

Amazon Leadership Principles

  • Moving Fast

 

So, Big Unicorns’ common goals,

  • Moving Fast
  • Focus on customers

 

Goal of DevOps

 

DevOps is not about a technology, DevOps is about a business problem.

 As above, unicorn companies have been applying DevOps methodology to respond customers’ needs in this fast moving business environment.

 

Steps of DevOps

 

Understanding the business value of DevOps is must do, but how about the implementation?  As to the steps of DevOps, the below will help you to approach your DevOps implementation; these can be boiled down to eight steps.

 

At first, there are four developmental steps as below.

 

No

Step

Contents

AutomationMethodology
1PlanUnderstand the problems and plan the solutions.No

Men use tools as aids to capture requests.

2

CodeCode software based on step 1.NoCoding by men varies productivity.
3BuildComplete step 2 coding to the degree of an executable implementation.Yes

Automation by CI/CD tools.

4

TestTest whether the completed implementation of the step 3 is running as planned.No
(Mostly No)

Partial automation is possible for the unit tests, API tests, and UI tests.

 

And the below are four operation steps.

As I said before, every company which apply DevOps make their efforts on how to minimize human intervention and automate above 8 steps.

 

5

ReleaseFix step 4 with the release.YesPrepare and generate the release through collaboration with the version control systems.
6DeployPrepare the deployment and service of step 5 depending on the operating system environment.Yes

Implement automation through a container or an orchestration tool.

7

Operate

Operation of the deployed service from step 6.

Yes/No

The goal is to operate automated processes with minimum human involvement.

8Monitor

Monitor service operation from step 7.

No
(Mostly No)

Identify anomalies of the system through monitoring log and event by human intuition and analytics.

 

Measuring DevOps and IT Performances

Next table shows how the unicorn companies go fast and do their own DevOps cycles forever. The faster cycle tends to be more fast product refinements and prompt user response.

 

  • Deploy frequency (Note: NOT delivery)
  • Mean Time to Recover (MTTR)
  • Lead Time for Changes

 

company

Deploy FrequencyDeploy Lead TimeReliabilityCustomer Responsiveness
Amazon23,000 / dayMinutesHigh

High

Facebook

1 / dayHoursHighHigh
Google5,500 / dayMinutesHigh

High

Netflix

500 / dayMinutesHighHigh
Twitter3 / weekHoursHigh

High

Typical enterprise

Once every 9 months

Months or quartersLow / Medium

Low / Medium

From a book, The Phoenix Project (2013)

 

The above chart shows that many companies run the DevOps cycle at least once a day, and the more the cycle they run, the more competitiveness they take.

 

Automation

Let’s take a look together with the automation, a keyword of the Forth Industrial Revolution and DevOps.

 

The needs for automation

To gain competitive advantage, what can be the best way to run the cycle faster?  The answer is automation.

 

Obstacles of Automation

Ideally, the fastest cycle performs from the complete automation especially in the fourth developmental step and the eighth operational step.  The trick is to strike a point of minimum human intervene and the maximum automation on development, testing, and deployment.

 

No

Group

StepContentsAutomationHow to automate

1

Development

PlanUnderstand the problems and plan solution method.No

Increase automated problem detection using AI.

2DevelopmentCodeCode software based on step 1.No

There is no complete automation in coding, but it is now an issue in AI field.

3

DevelopmentBuildExecute and implement coding from step 2.Yes
4DevelopmentTestTest the implementation of step 3.No
(Mostly No)

Check the error points.

5

OperationReleaseFix the release of step 4.Yes
6OperationDeployPrepare the deployment and service of step 5 to meet each operational environment.Yes
7

Operation

OperateRun the deployed service.Yes/NoIs the current service running normally?
8OperationMonitorMonitor various statuses while running the service of step 7.No
(Mostly No)

Is there any problem with the network, system, resources, etc.?

 

 

The key to the lead of competitiveness remains whether the resolution made it through automated testing, automated operations, and automated monitoring.

Most APM products need to insert some code at step 2 (code) and wait at step 7 (operate) and get the result at step 8 (monitor).

 

What ARGOS can do! 

 

ARGOS can make a difference in step 8.

No

Group

Step           ContentAutomationWhat ARGOS can do
1DevelopmentPlanUnderstand the problems and plan solution.No

2

DevelopmentCodeCode software based on step 1.No
3DevelopmentBuildTest the implementation of step 3.Yes

4

DevelopmentTestTest the implementation of step 3.No
(Mostly No)
ARGOS can automate UI scenarios and report the anomalies.
5OperationReleaseFix the release of step 4.Yes

6

Operation

DeployPrepare the deployment and service of step 5 to meet each operational environment.Yes
7OperationOperateRun the deployed service.Yes/No

ARGOS can run the service from the user perspective for a web or app, and analyze anomalies.

8

OperationMonitorMonitor various statuses while running the service of step 7.No
(Mostly No)

ARGOS specializes its service in GA or APM industry by providing the analytics based on the real user action while existing APM services are focusing just on monitoring service.

 

Argos can automate step 4 and 7 to reduce the number of cycles and the cycle time without any code level modification.

 


Jerry Chae
CAO, ARGOS LABS