ukstar 2017 london- parasoft

Test ing anyth ing, any t ime wi th conta iner ized serv ice v i r tua l izat ion Leveraging on-demand virtual test environments for earlier, better and faster testing

27-28 February 2017 LONDON

Rix Groenboom Parasoft

Paraso& Proprietary and Confiden1al 2

Who has an applica1on like calculator ?


Who has a process like this ?


Agenda

§ Modern (mes and need for Virtualiza(on § Virtualiza(on per phase

§ Development §  Test § Acceptance § Pre-‐produc(on

§ Requirements & implementa(on §  Func(onality (Dev) § Deployment (Ops)

§ Wrap up


Agenda




§ Wrap up


Once upon a 1me …

Business Logic

Data

Client A Client B q  Large applica(ons q  Few Development teams q  Few Test teams q  Few Opera(ons team q  Few releases q  Few Dependencies q  Life was easy … but slow

Batch processes


Modern 1mes

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP


Components: Dev -‐> Tst -‐> Acc -‐> Prod

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

D

T

A

P

D

T

A

P

D

T

A

P

D

T

A

P

D

T

A

P

D

T

A

P

D

T

A

P

D

T

A

P

D

T

A

P

D

T

A

P


DevOps teams

§ Many teams § Independent § Different cycle speed §  Test oVen §  Test early §  Test automa(cally

“Modern (mes”


Decoupling by virtualiza1on

§ Prevent teams slowing down by other teams § Prevent skipping tests in early stages § Extend test cases

§  Depends on stage in release cycle §  Various performance profiles §  Various deployment types §  Various transports (protocols) §  Various payloads (messages-‐types)


Agenda




§ Wrap up


Development

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

D

T

A

P


Development

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

D

T

A

P

Virtualize N,O


Development

B API

HTTP D

T

A

P

Virtualize N,O

§ Virtualize on local machine

§ Unit test § Func(onal tests § Fully decoupled § Fully independent


Test

Data NoSQL

A API

B API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

D

T

A

P C API


Test

Data NoSQL

A API

B API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

D

T

A

P C API

Virtualize P,X,Y,Z,

Third Party


Test

Data NoSQL

B API N

SVC

O SVC

Y BA

TIBCO

HTTP TIBCO

MQ

PROP

SOAP SOAP

JSON

PROP

PROP

PROP

D

T

A

P

Virtualize P,X,Y,Z,

Third Party

CD pipeline

Test


Test

§ Choose “cut-‐off” § Deploy AUT § Deploy Neigbours § Virtualize

§ On Server § Configured environment §  Supports custom transport §  Supports custom payload

§ Func(onal tests § Environment choice

§  Test chain

Data NoSQL

B API N

SVC

O SVC

Y BA

TIBCO

HTTP TIBCO

MQ

PROP

SOAP SOAP

JSON

PROP

PROP

PROP

D T

A P

Virtualize P,X,Y,Z,

Third Party

Test


Acceptance

§ Focus on non-‐func(onal requirements § Component central

§ DevOps team responsibility § Capacity § Stability § Performance § Resilience


Acceptance

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

D

T

A

P


Acceptance

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

D

T

A

P

Virtualize N,O

Test data


Acceptance

B API

TIBCO

HTTP SOAP

JSON

D

T

A

P

Virtualize N,O

CD pipeline

Test Test data

§ Reconfigure §  Slow response § No response

§ Virtualize § Reconfigurable § Performant §  Stable § Representa(ve data set


Pre-‐produc1on

§ Does not fit in pure DevOps § However, oVen (s(ll) mandatory § AKA “staging” § Produc(on like as possible

§ Sizing § Configura(on § Data

§ “Tomorrows produc(on”


Pre-‐produc1on

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Z BA

Y BA

Data RDBMS

Data Prop

Third Party SVC

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

P’

P

P

P

P

P

P

P P

P: Produc-on version P’: New version

P


Pre-‐produc1on

Data NoSQL

A API

B API

C API

M SVC

N SVC P

SVC

O SVC

X BA

Y BA

Data RDBMS

Data Prop

HTTP

TIBCO

HTTP TIBCO

HTTP

MQ

MQ

PROP

SOAP

SOAP

JSON

JSON

SOAP

PROP

PROP

PROP

P’

P

P

P

P

P

P P

P: Produc-on version P’: New version

Virtualize Z,

Third Party

Test

Test

Test

P


Agenda




§ Wrap up


Requirements gathering (1)

Based on these user-‐stories: § Local and Server deployment § Mul(ple Transport protocols § Mul(ple Message payloads § Flexible Data-‐set configura(on § Run-‐(me deployment § Run-‐(me data provisioning

Let’s make a “Separa(on of concerns”


Requirements gathering (2)

DEV (building Virtualiza(ons): § Message responders: XML, JSON, EDI, SWIFT, Text… § Logic: Synchronous, Async, Stateful, “batch” § Data: File, CSV, SQL, NoSQL OPS (running Virtualiza(ons): § Protocol listeners: HTTP, JMS, MQ TCP/IP, File, … § Deployments: Endpoints, Queues, JNDI § Performance profiles: Delay, Time Out, Errors

§  Some part of the Logic (func(onality) §  Some part of the Deployment (transport)


Deployment strategies (OPS)

Two main deployment strategies: § Central: Client / Server model

§ Repository containing Implementa(ons (PVA-‐files) § Environment Manager to control provisioning § Op(onal Load Balancing with mul(ple servers

§ De-‐central: Container or Cloud model § API driven launch of local servers § GIT based deployment of Implementa(on (PVA files)

Towards Infrastructure as code: Completely scripted construc(on


Central Deployment (client / server)

Virtualize SME

Environment Manager

Stage/Test Server

Production Server(s) Failover Server(s)

Desktop

Application Under Test


De-‐central Deployment (container based)

Paraso& Environment Manager


Towards scripted deployment

§ Infrastructure as code (IaC): § Docker & Java & Tomcat § Set protocol jars & Deploy Virtualize (war)

§ Sta(c implementa(on: § Deploy PVAs (model) from GIT § Provision test-‐data from a Data Repository

§ Dynamic implementa(on: § Build PVA files on the fly (using API) § Example:

§ Given “customer with 2 credit cards” § When “dele7ng one card” §  Then “one card should be le;”


Towards scripted deployment

B API

TIBCO

HTTP SOAP

JSON

D

T

A

P

Virtualize N,O

(deployment descrip(on)

CD pipeline

Test

Logic&Data N,O

§ Pipeline: § Deploy SUT (eg B API) § Deploy Virtualize (docker) § Ini(alize N,O placeholders

§ Test-‐execu(on: § Populate N, O (“Given”) § Execute (“when/then”) § (and repeat)


Deployment summary § Central server and central deployment:

§  “Waterfall”: Central “test data team”

§ Central server and de-‐central deployment: §  “Agile”: teams have control over Virtual Assets

§ De-‐central Server and central deployment §  For 3rd party interfaces / complex backends §  “Cer(fica(on environment”

§ De-‐central server and de-‐central deployment §  “DevOps”: Deploy Virtual Assets as part of pipeline §  Sta(c: “library of PVA’s” and provisioning §  Dynamic: “on the fly crea(on” (given-‐when-‐then)


Agenda




§ Wrap up


Wrap up

§ Virtualiza(on decouples your components § Decoupling oils the DevOps / CD pipeline § Virtualiza(on requires:

§ Various deployment models: Local, Server, Docker, Cloud § Ready to use transports and payloads § Extensibility where needed § Dynamic reconfigurable (provisioning) § Scalable to use in performance test § Integra(on in CD pipeline (APIs)

§ Like a Parrot / Turing Test:


Q & A

ukstar 2017 london- parasoft

Software