The document discusses how an SDDC case study assessed a company's technical architecture, processes, and people to reduce infrastructure costs by 50% within 3 years through migrating to a software-defined data center. It presents 3 technology stack options for implementing an SDDC and compares their considerations. The SDDC is projected to deliver dramatic and sustainable 75% cost reductions through server, storage, networking, and labor savings.
3. 3
Who Uses the SDDC Today?
Industry-standard hardware
Custom-built
Expensive
SDDC
Industry-standard hardware
Enterprise-class attributes
Open API set
So, What’s Different Now?
Commercially Available, Software-Defined Data Center
Who Uses the SDDC Today?
5. 5
Primary IT Business Driver
Reduce infrastructure OpEx costs
by at least 50% within 3 years
6. 6
Assessment Overview
• Technical architecture including network and storage
• Assess readiness of the enabling SDDC technologies
• Operating model and organizational structure
• Workflow processes and automation
• Cost model focusing on CapEx and OpEx savings
• Migration plan
Assessed
People – Process – Technology
7. 7
Management
Plane
Servers Network &
Security
Storage
Virtualized Network &
Security
Hypervisor Virtualized Storage
Application develop Application deploy Application run
PaaS
Monitoring and
analytics
Service
continuation
Service design
Service
automation
Infrastructure
Layer
Control
Plane
Application
Plane
ITaaS
Cloud Reference Architecture
IaaS
ITFM
Capacity
management
8. 8
Architectural Options Considered
Option Attributes
1 Converged vBlock, VCE, “cloud in the box”,etc.
2 SDDC Enabled Virtualized Network and Cloud-Aware Storage
3 Full SDDC Virtualized Network and Virtualized Storage
9. 9
Technology Stack Option 1: Converged
All the components of the stacks are tried and true and available today
Its architecture however unleashes the flexibility of cloud through administration automation and provisioning/release self service
Hardware Software
Management Plane
Application and data deployment service
Oracle, MsSQL, DB2, Sybase, SAS, Cognos, Websphere
Current Business Applications
Storage
virtualization
controller
Hypervisor
Virtual data center
Distributed vSwitch
Virtual firewall
Software layer 2 extension
Unified storage
Unified blades Core – Edge tier 1 switches
NAS (SATA,
SSD)
Converged 10GbE
Fiber channel
Application
Plane
Control
Plane
Infrastructure
Layer
Provisioning and
deployment management
Security management
Monitoring system
Financial management
system
Predictive capacity
management
Service continuation
management
Self-service portal
Management Plane
Key feature
10. 10
Technology Stack Option 2: SDDC Enabled
• The stack takes advantage of Software Defined Networking on top of a tried and true unified infrastructure:
Either unified blades on which the network control pane would be overridden by SDN
or self-built unified blades bundling unified storage to industry-standard servers according to the Customer’s specifications
Hardware Software
Application and data deployment service
Oracle, MsSQL, DB2, Sybase, SAS, Cognos, Websphere
Current Business Applications
Storage
virtualization
controller
Hypervisor
Virtual data center
Software defined network
Virtual firewall
Software layer 2 extension
Unified storage
Unified blades Industry Standard Switches
NAS (SATA,
SSD)
Converged 10GbE
Fiber channel
Application
Plane
Control
Plane
Infrastructure
Layer
Provisioning and
deployment management
Security management
Monitoring system
Financial management
system
Predictive capacity
management
Service continuation
management
Self-service portal
Management Plane
Key feature
11. 11
Technology Stack Option 3: SDDC
Approach will require highly skilled software defined data center resources, in particular experts in software defined
infrastructure configuration and monitoring
This architecture requires all workloads to be virtualized
Hardware SoftwareKey feature
Application and data deployment service
Oracle, MsSQL, DB2, Sybase, SAS, Cognos, Websphere
Current Business Applications
Hypervisor
based storage
virtualization
Hypervisor
Virtual data center
Software defined network
Virtual firewall
Software layer 2 extension
Industry
Standard
Servers
Industry Standard Switches
Locally attached
storage (SATA,
SSD)
Converged 10GbE
Application
Plane
Control
Plane
Infrastructure
Layer
Provisioning and
deployment management
Security management
Monitoring system
Financial management
system
Predictive capacity
management
Service continuation
management
Self-service portal
Management Plane
12. 12
Architecture Option 3: SDDC
Systems Management
Application Plane
- Application Provisioning
- Data Provisioning
- Cloud Aware Development
Management Plane
- Provisioning Portal
- Analytics
- Capacity Management
- Configuration Management
- Security
- Chargeback / ITFM
- Application Performance
Management
Control Plane
(Adapters)
- Automated Recovery
- Virtualized Storage
- Multi-tenancy Director
- Cloud Connectors
- Virtualized Networking
Hypervisor
vSwitches
SDS Datastore
Software Defined Storage (SDS) Layer
Whitebox Servers w/ DAS
Software-Defined Networking (SDN) Layer
Commodity Switches
vAPP
App
OS
App
OS
App
OS
vAPP
App
OS
App
OS
App
OS
vDC – Tenant 1
vAPP
App
OS
App
OS
App
OS
vAPP
App
OS
App
OS
App
OS
vDC – Tenant 2SDN Appliances
Nicira Controllers
Nicira Gateways
Virtual Appliances
Replication
vFirewall
vAntivirus
13. 13
Considerations by Architectural Option
Existing Target
Transition from multiple
heterogeneous facilities to fewer
homogeneous facilitiesSmall number of highly
utilized datacenters with
replication/mobility
Underutilized datacenters that
are designed for purpose
1 52 3 4
Reduce vendor lock-in and move to
using more industry standard
hardwareMassive over-provisioning of
expensive proprietary
infrastructure hardware
Dynamic / right-sized provisioning
of industry standard hardware
b Hardware 1 52 3 4
Greater use of automation and
service-oriented software shifts
hardware-specific engineers to be
more service oriented
Expensive and poorly scaled
labor structure
Lean silo-less org focused on
design (rather than hardware
config / management)
Labord 1 52 3 4
Implement full automation and
orchestration at the management
and services planeComplex set of management
and automation tools for
managing environment
Streamlined set of tools with
heavy automation at service
and infrastructure layer
Tools /
process
c 1 52 3 4
To evolve from applications on
proprietary platforms to applications
on generic x86 platforms may
require migration projects
a Datacenter
Legacy applications requiring high
cost hardware and labor support
Modern cloud-ready apps
only
Legacy
applications
e 1 2 3 54
X Converged X Full SDDCX SDDC Enabled
14. 14
ITIL Categories Remain but Underlying Processes Change
To…Process area
Configuration
management
▪ Real-time updates of frequently changing configurations
Capacity
management
▪ Predictive capacity management and auto-scaling app performance
Incident
management
▪ Skills shift towards running the critical software layer reliably
Performance
management
▪ Measure performance at the logical layer and for application SLAs
Physical resource
management
▪ Replace hardware with problems instead of troubleshooting
Request
fulfillment
▪ Self-service fulfillment from a service catalog with standard offerings
NOT EXHAUSTIVE
DATA CENTER OF THE FUTURE – OPERATING MODEL
15. 15
SDDC Organization Overview – Greenfield Org
Service design
and build
lead
Service
management
lead
Portfolio/catal
og manager
Relationship
manager
SLA/OLA
manager
Finance
manager
Procurement /
Vendor mana-
ger
Project
manager
Risk manager
Service
delivery lead
Enterprise
architect
Cloud
automation
lead
Cloud auto-
mation engi-
neer
Service
operations
lead
Lead solution
architect
Virtualization
architect
Security
architect
Storage
architect
Network
architect
Platform
architect
System
engineer
Incident
support lead
Incident
support
analyst
NOC analyst
Data center
lead
Data center
operator
Engineering
lead
System
admins
Platform
admins
Security
admins
Network
engineer
System
analyst
Platform
analyst
Service
organization lead
1-1
3-6 3-6
2-2 2-2
2-4 4-6
2-2
1-1
2-3
1-1
2-4
24-30 6-9
2-3
50-70
8-12 8-12
8-12 8-12
8-12
1
1-1
1-2
4-6
2-3
4-6 4-6
4-6 4-6
4-6 8-12
1-4
16. 16
SDDC Organization Overview – Corporate Integrated
Security
admin
Tenant Ops
lead
Portfolio
manager
Service
design, dev, &
release
lead
Service
Owner
Relationship
manager
Service
architect
Service
quality mgr
Infrastructure
Ops lead
Infra design
and build
lead
Enterprise
architect
Cloud
developer
lead
Cloud
developer
Infrastructure
operations
lead
Infra architect
& engineering
lead
Cloud
architect
Security
architect
Network
architect
Platform
architect
Cloud
admin
Storage
admin
Performance
analyst
Network
admin
Compliance
analyst
Cloud Ops
organization
lead 1-1
1-1
X-X
1-1
X-X
X-X
X-X
X-X
1-1
1-1
X-X X-X
X-X
1-1
X-X
X-X X-X
Storage
architect
X-X X-X
X-X
X-X
X-X X-X
X-X X-X
X-X
Cloud Infrastructure Ops
Capacity
analyst
X-X
Service
Developer
X-X
Service
operations
lead
Service
capacity
analyst
Service
administrator
1-1
X-X
X-X
Service
performance
analyst X-X
Tenant Ops
External to Cloud Ops
17. 17
Baseline Converged SDDC Enabled SDDC
Option Considered
Estimated TCO Reduction by Option*
2013 vs 2016
Financial Analysis – Dramatic Sustainable Cost Reduction
100%
- 38%
- 54%
- 75%
Costs include:
• Annual Depreciation
• Labor Spend
• Data Center Costs
• Data Transmission Costs
• Legacy Costs Carried Forward
* Actual results may vary by organization
18. 18
Dramatic Sustainable Cost Reduction
SDDC is projected to deliver 75% in steady-state IT infrastructure savings*
* Actual results may vary by organization
17%
12%
28%
37%
6%
Savings Contribution
Servers Storage Networking Labor Other
19. 19
$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
2014 2015 2016 2017 2018
Cost per VM (baseline)
Cost per VM (baseline/standup/migration)
Cost per VM (baseline/standup/migration/writeoffs/severance/retention)
Workload Costs Reduce Over Time
Cumulative #
of available
workloads
~31K~8K ~64K ~93K ~115K
Cost per virtual machine will experience a significant drop
over time with increased scale and maturity of processes
Approximate Cost of
Amazon/Google per VM
20. 20
Top 5 Value Drivers – Why Is SDDC Attractive Now?
Dramatic Sustainable Cost Reduction
• Realize 40-75% cost reduction at full SDDC: ~75% CapEx / ~56% OpEx reductions
Organizational Efficiencies
• “New” skillsets and headcount optimizations are required and allow much more to be done by
much fewer, more cross-functional staff.
Datacenter Agility and Efficiency
• High levels of automation and self-service drive a much more agile datacenter environment and
greatly reduce time to market of services and applications.
Operational Efficiencies
• Simplified, consolidated, and heterogeneous management and orchestration toolsets allow for a
fully-realized service-oriented organization.
Technology Stack Maturity
• The core software-defined technology stack is here today and will get better with time. Today we
leverage 3rd party vendors to support SDDC. Fully-realized SDDC solution by end of year.
22. 22
Two Paths
Legacy Systems
Rely on
infrastructure
stack to provide
SDDC
abstraction
Migration without
application re-
development
IaaS offering
New Development
Have intelligence
built into the
applications to
abstract hardware
layer
Applications must
developed in a
modern application
framework
PaaS offering
28. 28
What Customers Want
Complexity is often required to expose simplicity
A cloud computing infrastructure has complexity in the form of technology and
automation that simplifies the experience for end-users.
Cloud Provider Organization or User
What Customers Want!
30. 30
Accelerate/PSO ELA Sprint Workshop and Demo Lab
Building hardware stack in Wenatchee
datacenter
Utilizing “VMware-friendly” vendors
• Hyve, Juniper, EMC
Multi-purpose hardware stack
Designed to show customers complete
kits they can build to meet their needs
Demonstrations and reference
architectures built on business
use-cases
• Utilizing SDDC to provide IaaS
• Utilizing SDDC to provide PaaS
• Utilizing SDDC to provide ITaaS
LEGEND:
= QSFP+ 40Gbps
= SFP+ 10Gbps
= RJ45 1Gbps UPLINK
= RJ45 1Gbps MGMT
= RJ45 1Gbps STORAGE
32. 32
IaaS Workflow
? Small System
Medium
System
Large System
Custom
System
$8460.00/year $9400.00/year $11,280.00/year TBD by ProjectSelf Service Portal
Cloud Automation
Base OS Builds
Blueprint Definitions
OS-level Compliance
Options
Standard Hardware Stack Medium Hardware Stack Extreme Hardware Stack Custom Hardware Stack
Resource Pools
OrchestrationEngine
33. 33
Infrastructure as a Service
Service Description:
We provide configuration, provisioning, operation, management and optimization of a secure,
undisturbed system environment for hosting a wide variety of customer applications.
Service Offering:
• Virtual Server Configurations:
• Standard X CPUs and X Memory
• High Performance X CPUs and X Memory
• Extreme Performance X CPUs and X Memory
• Custom Configuration Available upon request
• Processor hosting management with the operating system, storage, connectivity and
infrastructure tools to optimize and support the environment end-to-end
• Preventive maintenance, protection, and lifecycle refresh
• Regulatory compliance support
• Includes secure, network connectivity and data storage and backup (up to 200 GB per
Instance)
• Snapshots every X hours with X versions, retained for X days
• 24x7 monitoring for availability and incident/problem management
Performance Targets:
Chargeback Basis:
• Standard: $8,460 per OS Instance
• High Performance: $9,400 per OS
Instance
• Extreme Performance: $11,280 per
OS Instance
• Additional Fee for Tier Availability:
• Tier 1: $1,500 per OS Instance
• Tier 2: $1,000 per OS Instance
• Tier 3: $600 per OS Instance
• Tier 4: $500 per OS Instance
• Tier 5: $300 per OS Instance
• Tier 6: $100 per OS Instance
Custom configurations priced upon
request
Cost Saving Tips:
• Consolidate and standardize on our
supported server configurations when
possible
• Plan changes and requests to
minimize expedited or emergency
changes
Additional Information:
• Service Delivery Owner:
• Service Desk Contact:
Service Level Service Level Description Service Level Target
Response to Provisioning
Amount of elapsed time from submittal of provision
request to operational status
1 Week
Incident Management
Target time in which staff will respond to incidents
recorded in incident management system
< 1 hour, 80% of the time
Virtual Server Infrastructure
Availability
The percent of time that the virtual server infrastructure
is available for normal business operations
Tier 1: 99.99%
Tier 2: 99.9%
Tier 3: 99%
Tier 4: 98.5%
Tier 5: 95%
Tier 6: <90%
38. 38
Other VMware Activities Related to This Session
HOL:
HOL-SDC-1313
vCloud Suite Use Cases - Infrastructure Provisioning (IaaS)
Group Discussions:
VSVC1006-GD
vCloud Suite and SDDC with Tom Stephens