This document discusses building a service provider cloud offering. It covers topics such as choosing between public, private and hybrid cloud models, selecting infrastructure including enterprise hardware or whitebox solutions, designing for efficiency and resiliency as a service provider, automating operations, and differentiating a cloud offering through value-added services or bundles. The document provides advice on starting small and working with partners if building your own cloud platform seems too complex.
Building a Service Provider Cloud Offering - MVMUG Sept2013
1.
Building
a
Service
Provider
Cloud
Offering
Arron
Stebbing
Cloud
Architect/Opera:ons
Manager
2. Who
Am
I?
• vExpert
2013
• vChampion
2011/2012/2013
• Work
History
– ZeKaGrid
(Public
Cloud)
– FlexScale
an
Empired
Offering
(Hybrid
Cloud)
– Thomas
Duryea
Cloud
(Hybrid
Cloud)
3. Why
build
a
Cloud?
•
•
•
•
•
On
Demand
Self
Service
Broad
Network
Access
Resource
Pooling
Rapid
Elas:city
Measured
Service
(Showback
/
Chargeback)
NIST
Defini:on
for
Cloud
Compu:ng
7. Hybrid
Cloud
Hybrid
Cloud
•
•
•
•
Networking
is
key
Business
Agility
Design
a
must!
OpEx
8. No
Cloud?
No
Cloud?
•
•
•
•
Is
this
Cloud
thing
really
worth
while?
What
will
it
mean
to
my
business?
Is
my
job
at
risk?
What
will
happen
if
I
don’t
go
cloud?
10. Pros
and
Cons
Enterprise
Hardware
Pros
• Global
Support
and
SLA’s
• More
features
to
upsell
• Transferable
Knowledge
Cons
• Cost
• Need
to
capitalize
on
upsell
11. “Enterprise”
Mindset
• Day
to
Day
solu:on
offering
• Reference
Architectures?
• Designed
for
Medium
to
Enterprise
Environments
• Vendor
Rela:onships
13. Star:ng
the
Journey
• Ensure
you
have
Full
Time
resources
• Review
and
Confirm
Bill
of
Materials
– Order
new/more
if
possible!
• Provide
High
Level
Design
• Detailed
Design
–
Reproducible
• Content
Management
System
– Wiki/IPAM/CMDB
14. Cost
Modeling
•
•
•
•
Cost
Recovery
Service
Level
Agreements
Service
Level
Objec:ves
Research
&
Development
15. Running
a
Cloud
• Automa:on/Orchestra:on
• Cross
Skilling
Required
– Compute/Storage/Network/Support
• On
Call
Rosters
• Infrastructure
Efficiency
• Monitoring/Aler:ng
– Informa:on
is
crucial
16. Right
Sizing
&
Efficiency
for
the
Cloud
•
•
•
•
Ignore
Defaults
Start
Small,
Upscale
Later
Workload
Based
Cluster
&
Datastores
Don’t
be
scared!
17. Help
is
Available
•
•
•
•
Leverage
Compe::on
Leverage
Vendors
Leverage
Reference
Architectures
Leverage
Community
– Blogs
– TwiKer