Download - The cloud reality facing traditional operations engineers

Transcript

The Cloud Reality Facing Traditional Operations Engineers

Gina Rosenthal

Product Marketing Manager, VMware

Who am I?• Product Marketing Manager –

currently VMware

• Past lives: EMC, Spanning, Dell, Inktank, linux sysadmin

• Live in Austin via Mass (from Florida)

• I like the beach, good stories and my dog Fred

• Connect/complain: @gminks

agenda

• Can the stuff you already know translate to managing cloud environments?

• Can you use on-premises tools to manage cloud environments?

• How can DevOps methodologies blend with ITIL/ITSM methodologies?

Wannacry virus

Why don’t you patch?

Thread https://twitter.com/wendynather/status/863361400890163200

1958

high availability

Users

Data

Application

App Admin

OS

Virtualization

Hardware

Network

On-Premises

backupsData

Application OS

Virtualization

disaster recovery

Hardware Network

Security

Protect it all

Monitoring

Understand what to do when

Users

Data

Application

App Admin

OS

Virtualization

Hardware

Network

On-Premises

backupsData

Application OS

Virtualization

disaster recovery

Hardware Network

Security

Protect it all

Monitoring

Understand what to do when

Data Center Hygiene

Have we learned from past innovations?

Mainframes Client/Server SAN/NAS Virtualization Containers Cloud

Doing a backup before an upgrade

hardware jumper password recovery

on a Cisco AGS

burning server images to a cd

How to manage memory via pointers in software

Hayes AT command set

formatting a floppy disk

loading the punch cards

ghost

Iron mountain RAID card drives on

floppy Floppy drives in

servers

Creating new re-write rules in sendmail.cf

Where the monitor and keyboard are

in rackSymmetrix inlines Dial up Floppy disks

The proper way to load a 9-track tape

Tweaking autoexec.bat and

config.sys

Drilling a hole into a floppy to double

capacityIPX/SPX

Great things don’t just appear from nothing….

Where did virtualization come from?An extremely abbreviated history

Mainframes. • 1st stable release of

IBM CP/CMS OS for CP-67

1972

Ability to run more than one program at a time (time sharing)

1st DOS emulator for UNIX• SoftPC

• Made software portable

• Run your apps on UNIX or DOS

1987

Cost savings, operational flexibility

VMware workstation• Provided ability to run

a hypervisor-based OS along side a client machine’s OS

1999

Cost savings, operational flexibility

Server virtualization• VMware released ESX

• Virtualization market leader

2001

Cost savings, operational flexibility

Where did containers come from?An extremely abbreviated history

Vocabulary break!

Virtualize at the HW level

Abstract the OS from the HW

HypervisorVirtualize at the OS level

Abstract the app from the OS

Containers

Mainframes • IBM MVS onwards have

had container technology

1974(ish)

chroot• Changes apparent root

system for a current running process

• Isolates processes to a specific location

1979

Security

Free BSD Jails • Each jail has its own IP

address

2001

Improved security, increased access to underlying OS

Solaris containers• Segregates file systems,

and combines resource controls into “zones”

• Similar to jails, adds cloning

2004

Adds management features

LXC (Linux Containers)Containers that interact with namespaces and cgroups.

• namespaces: provides segregation (exp: chroot=FS segregation)

• cgroups: limits the amount of resources a process can use

2008

Adds management features

LMCTFY (Google containers) • Open Source version of

Google’s container stack (Linux application containers).

• Abstracted cgroupsupport, better QoS

2013

Adds management features

Docker• Open-source project

that automates the deployment of apps inside containers.

• Makes containers easier to use and deploy

2013

Adds automation, deployment management

You already understand the basics• Containers virtualize files at the OS level vs the hardware

level• Same idea, different components

• Have you ever…• chrooted a service?

• You understand one component of a container. • Managed jumpstart or kickstart?

• You understand the basics of automation• Managed Linux in production?

• You should understand kernel basics, which is what all of this is built on!

We manage appsOps builds that infrastructure that hosts apps.

Ops is responsible for the data created by those apps.

Containers are only the tip of the next wave of innovation

What is a cloud?

Cloud Deployment ModelsUsers

Data

Application

OS

Virtualization

Hardware

Network

Users

Data

Application

App Admin

OS

Virtualization

Hardware

Network

On-Premises SaaS

Users

Data

Application

App Admin

OS

Virtualization

Hardware

Network

IaaS

Users

Data

Application

OS

Virtualization

Hardware

Network

PaaS

App Admin App Admin

Users

Data

Application

App Admin

OS

Virtualization

Hardware

Network

On-Premises

backupsData

Application OS

Virtualization

disaster recovery

Hardware Network

Security

Protect it all

Monitoring

Understand what to do when

Data Center Hygiene

Plan cloud deployments like any other data center change• How will you secure it?

• Most likely, you’ll depend on your cloud provider. • How do you know they’ll secure it?

• What about backups?• Will the cloud provider’s backup plan satisfy your regulatory

obligation?• Do you need a secondary backup solution?

• How will you know what’s going on?• What sort of monitoring does the cloud provider supply?• Should you use a secondary monitoring solution?

Does it blend?DevOps + ITSM == ?

What is devopsanyways?“the practice of operations and development engineers participating together in the entire service lifecycle, from design through the development process to production support”

-The Agile Admin

This is all great but….I have legacy infrastructure, apps, and processes to support.

And how can we fit this into our current processes?

Put down the scotch…

Devops ITSM

Agility Discipline

Site Reliability Engineering (SRE)

• DevOps is a methodology, but Site Reliability Engineering (SRE) is a new job role

• Championed by Google, SRE is the operational discipline required to continuously develop the infrastructure devs need

• Sysadmins assemble existing software components and deploy them to work together as a service.• Via O’Reilly’s Site Reliability Engineering

Get your learn on now!

Remember our shared heritage.

Connect with me!

• Twitter is best! @gminks

• I blog about tech things here:https://24x7itconnection.com/category/gina-rosenthal/