Virtualization Challenges
1.
Application performance
2.
Security
3.
VM sprawl
4.
Licensing costs
5.
Stuck on storage
6.
Virtual roadblocks & Silos
Virtualization Challenges
1.
Application performance
– Virtualization does not add any significant performance issue – Resource saturation may affect multiple application and services – applications aren’t yet tuned for virtual environments
• unused resources are not freed-up • virtualization new features not used
2.
Security
– Compromise of the Virtualization Layer
– Lack of Visibility and Controls on Internal Virtual Networks – Workloads of Different Trust Levels
– Adequate Controls on Administrative Access
Virtualization Challenges
3.
VM sprawl
– Number of OS instances is increasing
– Licensing, compliance, administration issues mount – Virtual machine lifecycle management (VMLM)
– Transition to Private Cloud solves most of these problems
4.
Licensing costs
– Hypervisor Cost
Virtualization Challenges
5.
Stuck on storage
– Storage I/O has been a limiting factor for large environments – Read / Write IOPS pattern has changed dramatically:
• Typical distribution read/write operations: 80/20
• Virtualization distribution read/write operations: 50/50 or even 40/60
– desktop virtualization produces new demands for I/O, especially write I/O
6.
Virtual roadblocks & silos
– Multiple generations of hardware
– Traditional design limitations result to silos of H/W
7.
Virtual Mobility Impacts Network Optimization
– Network Virtualization is following System’s virtualization
BUSINESS NEEDS
TECHNOLOGY NEEDS
Speed: Want/Need something ASAP –
“Cloud is quick” Cost: Must be low cost with minimal/no upfront investment Fits Business Model: Must be able
to fit the way we do business
Ease of use: Must be easy to use and adopt
Must be stable: reliable application and delivery
infrastructure
Must be secure: Must be able to protect my data!
Must be able to integrate with other business applications
Organizations Wrestle With Utilizing the Cloud to Solve Issues
Pressures to Move to the Cloud
IT
Many workloads are migrating to public or private cloud “camps”
Workloads shifting to Cloud
TRADITIONAL IT
• Server Capacity On Demand • IT Management
• Business Apps (CRM, ERP)
• Personal Productivity Apps • Website Creation & Management • Storage Capacity on Demand • App Dev. & Test
• Tech. Computing Apps • Data Analysis and Mining
• Custom Applications
• Applications with sensitive data
PRIVATE CLOUD1 PUBLIC CLOUD1 • IT Helpdesk
• Collaborative Applications • Data backup/Archive
Cloud Service Layers
Cloud Infrastructure Services (IaaS)
Cloud Platform Services (PaaS)
Cloud Computing – characteristics
Pay as you go. Business customers only pay for products and services actually used on a metered, chargeback basis under flexible service agreements, as opposed to fixed term contracts
On demand. Requests for IT services should be delivered immediately as requested, supported by real time processing, to cut the time required for the fulfillment process
Automation. Business customers must be able to request services from a catalog via self service interfaces
Standardization. The IT organization defines and enforces the use of a set of standard services articulated in a comprehensive service catalog, encouraging reuse and achieving cost predictability Workload portability. IT services are designed to be fully portable between infrastructures and
Best practices
Drive organization change from the top down
Develop clear cloud service quality objectives and metrics
Build a holistic cloud operating model
Enforce standardization to take advantage of cloud
The results: Savings, speed, agility and a
platform for the future
Early cloud returns fund new strategic IT initiatives
Cloud savings drive down the amount spent on basic IT
infrastructure operations
Workloads are deployed to the most efficient internal or external
locations
Reduced wait times for infrastructure save additional time and
IT turns business “on,” instantly
The future
• Everything & everyone is connected
• Everyone expects immediate gratification & instant results
• Business & IT are one and the same
How to move forward?
• Clear logical separation of systems
• About 80% of systems are not business critical and/or can be standardized
• Only about 20% are business critical and cannot be standardized
• Create categories of systems
• Target for non-business critical systems that can be standardized: Public Cloud
• Target for business-critical systems that can be standardized: Private Cloud
• Target for systems that cannot be standardized: Legacy Infrastructure
• Disentangle physical systems to enable migration to future platform
• Establish clear interfaces, SLAs, security and performance requirements
• Move cloud-target systems towards SOA-enabled applications
• Align organization to new operating model
Ideal State for cloud computing
Streamlined, cooperating, shared services
Standard APIs and Data Models
Scale up - Scale down Capacity
Movable workloads – affinity with market access, services
and data location
Wide choice of:
Data sources