1. Cloud Computing
J Srinivasa Rao, Co-ordinator
Skill Development Center
Department of Technical Education
Government of Andhra Pradesh
cloud@sdcdte.org
2. 2
virtual datacenter OS – The Next Gen of Virtualization
VMware Infrastructure is extended to a virtual datacenter OS in 2009
A virtual datacenter OS:
Aggregates server, storage and network hardware into a shared resource or ‘internal cloud’
Allocates this shared resource among applications precisely and efficiently
Provides built in services to ALL applications such as availability, security and performance scalability
Federated with external clouds so cloud computing is easily accessible for enterprises
VDC-OS is to the entire datacenter what Windows and Linux are to a single server
4. 4
Key Industry Trends
Innovation with new architectures
Hard to ensure
consistent service levels
IIS .Net Oracle
Apache
Web
sphere MySQL
More x86 in datacentersIncreasing cores and memory in servers
Abundant cloud capacity on demand Limited , complex
access
8. virtual datacenter OS from VMware
vCloud
vCenter
On-premise Infrastructure
SaaSLinux GridWindows J2EE.Net
VMware Infrastructure -> virtual datacenter OS
Application
vServices
Scalability
Infrastructure
vServices
SecurityAvailability
vNetworkvStoragevCompute
Cloud
vServices
…….
Web 2.0
VDC-OS is to the entire datacenter what Windows and Linux are to a single server
9. VMware Infrastructure -> virtual datacenter OS
App
OS
App
OS
App
OS
The Evolution of VMware
Infrastructure
Application
vServices
Scalability
Infrastructure
vServices
SecurityAvailability
vNetworkvStoragevCompute vCloud
Off-premise
Cloud
ESXvCenter
Application
Management
Infrastructure
Management
DesktopLinuxSaaS Windows J2EE.Net
10. 11
“VMware already has had transformative
impact on how we do things: our server farms are
aggregated into clusters that act as a single large
computer that can guarantee service levels to
applications. We no longer worry about scheduling
downtime for hardware maintenance, or worry about
hardware failures. And all of that delivered while
reducing the infrastructure cost per application by
more than 50%.
VMware Infrastructure as a VDC-OS
Hill AFB
11. 12
*Source: IDC and VMware TAM program
Infrastructure Cost per App
$14,235
$5,694
Before VMware After VMware
The VMware Effect: Customer
Breakthroughs
60% Reduction
in Cost
2–3x Gain in
Productivity
Workloads per Admin
30–75
Before VMware After VMware
100–250
12. virtual datacenter OS from VMware
Off-premise
Cloud
vCenter
On-premise Infrastructure
SaaSLinux GridWindows J2EE.Net
VMware Infrastructure -> virtual datacenter OS
Application
vServices
Scalability
Infrastructure
vServices
SecurityAvailability
vNetworkvStoragevCompute
Cloud
vServices
…….
Web 2.0
13. New Application vServices for the best place to run all applications
CURRENTNEW
VMware Infrastructure -> virtual datacenter OS
Application
vServices
ScalabilitySecurityAvailability …….
•VMware Fault
Tolerance
•vCenter Data
Recovery
•VMware VMsafe
• IBM, McAfee, Checkpoint,
Radware announce VMsafe
products
•Hot add of virtual
CPU, memory and
devices
•Very large virtual
machines with 8-
virtual CPUs and
256 GB of RAM
• HA, VMotion,
Storage VMotion,
NIC/HBA teaming
provide resiliency to
downtime
• ESXi 32 MB of code,
locked down interfaces,
no general purpose OS
dependence
• DRS shares and
reservations allow
apps to shrink and
grow based on priority
14. 15
HA
VCB
NIC & HBA Teaming
VMotion
Storage VMotion
Network Redundancy
VMware Solutions Maximize
Uptime
PerformancePlanned Downtime Unplanned Downtime
VM Failure Monitoring
Virtual Machines
Server
ESX Server
App
OS
App
OS
App
OS
App
OS
App
OS
Storage
Interconnect
Site Recovery Manager
Availability
15. 16
New Solutions for Reduced
Downtime
Server
ESX Server
Storage
App
OS
App
OS
App
OS
Zero downtime, zero data
loss continuous availability
Fault Tolerance
Integrated backup and
recovery appliance
Data Recovery
Availability
16. 17
App
OS
App
OS
App
OSXX
App
OS
App
OS
App
OS
App
OS
X
Zero downtime, zero data loss
No complex clustering or specialized
hardware required
Single common mechanism for all
applications and OS-es
Single identical VMs running in lockstep
on separate hosts
Zero downtime, zero data loss failover
for all virtual machines in case of
hardware failures
Integrated with VMware HA/DRS
VMware ESX VMware ESX
VMware Fault
Tolerance
FTHAHA
2009
Availability
19. 20
The Hypervisor is the Foundation
Partitions a server into virtual machines
Reduces hardware, power, and cooling
with the performance and features of ESX
Plug-and-Play
Minimal configuration. Run VMs in minutes
Integrated in server hardware
OS-Independent, thin, 32MB architecture
Unparalleled security and reliability
ESXi is the next generation of the market-leading ESX hypervisor
VMware ESXi
Virtual Machines
App
OS
App
OS
App
OS
App
OS
Security
20. 21
Application
Operating System
Protection
Engine
VMware Infrastructure
VMware VMsafe
API that enables protection of VMs
by inspection of virtual components
in conjunction with hypervisor
Isolation of protection engine from
malware
Broad ranging coverage of virtual
machine CPU, memory, storage
and network
Security
21. 22
Ecosystem Enablement with VMware
VMsafe
2009
Multi-function Security Appliance
VMware ESX
A
pp
OS
App
OS
App
OS
App
OS
App
OS
Security
VM
vNetwork Distributed Switch
VMware ESX
A
pp
OS
Security
VM
Integrated, more effective,
comprehensive security solutions
within the virtual infrastructure
Better security than physical servers!
Agent-less deployment of partner
security services
Single security VM for multiple
security services AV, Firewall, IPS
Security policy and state moves
with virtual machine
Security
22. 24
64 GB
4 CPUs
App
OS
256 GB
8 CPUs
Scale Out Applications for Assured
QoS
Zero downtime scale out of
virtual machines
Scalable virtual machines
Hot add of
CPU
Memory
PCIe devices
2009
Scalability
App
OS
23. virtual datacenter OS from VMware
Off-premise
Cloud
vCenter
On-premise Infrastructure
SaaSLinux GridWindows J2EE.Net
VMware Infrastructure -> virtual datacenter OS
Application
vServices
Scalability
Infrastructure
vServices
SecurityAvailability
vNetworkvStoragevCompute
Cloud
vServices
…….
Web 2.0
24. Infrastructure vServices and Cloud vServices
VMware Infrastructure -> virtual datacenter OS
Infrastructure
vServices
vNetworkvStoragevCompute Cloud
vServices
•vStorage
Thin
Provisioning
and Linked
clones
•vNetwork
Distributed
Switch
•Third party
virtual
switches
•CPU/Memory
optimization
•DRS
•VMware vCloud
•Network
VMotion
•VMDirectPath
•Paravirt SCSI
•vStorage
VMFS
•vNetwork
Offload
technologies
•VMotion
•Storage
VMotion
CURRENTNEW
Lowest TCO through maximum efficiency
25. 28
2009
Storage
Networking
Virtual Machines
CPU
Memory
Highest Single Server Resource Efficiency
128 cores and 512 GB of
physical RAM
Hardware Scale Out
Lowest CPU overhead
HW Assist
Purpose Built Scheduler
Maximum memory
efficiency
HW Assist
Page Sharing
Ballooning
Wirespeed (9Gb today)
network access
VMDirectPath
Offload
Greater than 200k iops
per second
Lower than 2ms latency
VMDirectPath
Para-virtualized SCSI
ESX Server
App
OS
App
OS
App
OS
App
OS
App
OS
8-way vSMP and 256 GB
of RAM per VM
Virtual hardware scale out
Virtual hardware
scale out
vCompute
26. 29
0
1000
2000
3000
4000
5000
1 2 3 4 5 6
ESX Maintains Performance During
Consolidation
Number of Virtual Machines or CPUs
SPECweb2005AggregateMetric
Native Scaling
Virtual Scaling
vCompute
28. 32
Aggregated view of virtual networking
Datacenter level networking
(versus host level)
Historical statistics follow the VM
A unified infrastructure for networking
services (monitoring, filtering, mgmt
via PVLANs)
Simplified setup and change;
seamless addition of capacity
Easy troubleshooting, monitoring
and debugging
Enables new security services
2009
vSwitch
vNetwork Distributed Switch
vSwitch vSwitch
App
OS
App
OS
App
OS
App
OS
App
OS
vNetwork Distributed Switch
vNetwork
29. 33
App
OS
App
OS
App
OS
App
OS
App
OS
Enterprise networking vendors can
provide proprietary networking
features in a VMware environment
Enables networking solutions to
monitor, control and manage virtual
networks
Networking/security solutions can
understand/ be aware of mobile,
dynamic virtual infrastructure
Simplicity and transparency for
network administrators
Unified management framework for
physical and virtual networks
Ecosystem Enablement 2009
Third Party Virtual Switch
vNetwork
31. 35
Significantly improve storage
utilization
Eliminate need to over-provision
virtual disks
Reduce storage costs by up to 50%
App
OS
App
OS
App
OS
Virtual machine disks consume only
the amount of physical space in use
Virtual machine sees full logical
disk size at all times
Full reporting and alerting on
allocation and consumption
vStorage Thin Provisioning
Physical
Storage
Virtual
Disks
10GB
20GB 40GB 100GB
10GB20GB 10GB40GB
ESX
30GB70GB
vStorage
32. 36
vStorage linked clones
Multiple virtual machines share
common base disk
Each virtual machine has own disk
that stores its writes to disk
Patches applies to base disk are seen
by all linked clones
Reduce storage costs for Virtual
Desktop Infrastructure by up to
90%
Improve storage utilization
Simplify patch process
OS
App
OS
App
Base
Disk
OS
App
OS
App
vStorage
33. 37
Delivers storage capabilities as
virtual appliances
Validated by VMware Ready Virtual
Appliance certification program
App
OS
vStorage Virtual Appliances
Simplify deployment and administration
Reduce obstacles to delivering full
virtualization experience
Easy transition to hardware-based
functionality as environment grows
ESX ESX
SVA SVA
Data protection
Availability
Storage management
App
OS
App
OS
App
OS
vStorage
34. 38
App
OS
App
OS
App
OS
App
OS
App
OS
vStorage API’s
Enhance integration of VMware
Infrastructure and storage partner
capabilities
Provide storage management tools
with visibility to virtual machines’
use of storage
Leverage array capabilities at
per-VM level
Fully utilize investments in
advanced storage capabilities
Simplify storage management for
virtual environment
VWware Infrastructure
Partner Storage and Management
vStorage
35. 39
How Do We Define The Cloud?
Cloud Computing according to
VMware
Lightweight entry/exit service
acquisition model
Consumption based pricing
Accessible over the internet
Scalable and elastic
Improved economics due to shared
infrastructure and elasticity
Cloud computing comes into focus only when you think about… a
way to increase capacity or add capabilities on the fly without
investing in new infrastructure, training new personnel, or licensing
new software. Cloud computing encompasses any subscription-
based or pay-per-use service that, in real time… extends IT's existing
capabilities.
“
“
Cloud vServices
37. 41
vApp – New Model for Describing and Deploying
Applications
Availability =
99.99%
Security = High
Performance =
500 msec
SLA Definitions
vApp
App
OS
App
OS
App
OS
VMware Infrastructure -> virtual datacenter OS
Application
vServices ScalabilitySecurityAvailability
Allows management of multi-tier
applications as a single entity
Utilizes industry standard OVF to
provide instructions on how to deploy
Templates, Clone and other operations
execute at the vService level
Simpler, application centric view of
management
Easier portability of applications
Applications can now be written to
monitor and scale themselves
38. 42
virtual datacenter OS from VMware
Off-premise
Cloud
vCenter
On-premise Infrastructure
SaaSLinux GridWindows J2EE.Net
VMware Infrastructure -> virtual datacenter OS
Application
vServices
Scalability
Infrastructure
vServices
SecurityAvailability
vNetworkvStoragevCompute
Cloud
vServices
…….
Web 2.0
41. 45
Simplified Management – What’s New
vCenter
Applications
Infrastructure
Quality of Service Management for Applications
vCenter AppSpeed: real-time performance monitoring
and remediation guarantee service levels
Proactive Management of Virtual Infrastructure
vCenter Orchestrator automates repetitive IT
processes with a robust workflow engine
vCenter CapacityIQ provides proactive capacity
planning for VMs, clusters and entire VI environment
vCenter Chargeback automates chargeback modeling
for greater visibility and control of IT costs
vCenter ConfigControl automates discovery and
tracking of configuration items for better visibility,
compliance and remediation
Host Profiles simplify setup and ensure standardization
43. 47
Automated
Provisioning
Automate the provisioning process
and VM lifecycle
Standardize how VMs are brought into
the infrastructure
Improve visibility & control over VM
lifecycle from cradle to grave
Reclamation of unused capacity with
policies to decommission idle VMs
Integration with multi-component
vServices based on OVF standard
Intelligent, policy-based deployment
Defined thresholds that consider
available resources before provisioning
API for third-party integration
vCenter UI integration
Provision DecommissionDeploy Track
Lifecycle Manager
Task
Management
Infrastructure Management
44. 48
Automate chargeback modeling
based on resource utilization
Improve decision making with end
user visibility to cost of service levels
Simplify management, tracking and
control over IT service costs
Costs based on service levels,
underlying infrastructure and custom
definitions of fixed costs
Chargeback tied to organizational
structure
Integrated with vCenter
Transparent Chargeback
Resource
Usage
Reporting
Costing
Model
Chargeback
Engine
vCenter
$
vCenter Chargeback
2009
Infrastructure Management
45. 49
Visibility into capacity and utilization
What-if analyses simulate effects of
capacity changes
Forecasts of capacity shortfall
Integrated with vCenter
Align capacity with business
demand
Reduce costs by reclaiming
unused capacity
Predict capacity shortages
proactively
Intelligent
Capacity
vCenter CapacityIQ
2009
Infrastructure Management
46. 50
Simplify physical host configuration,
administration and compliance
Automates discovery, tracking and
dynamic search / reporting / modeling of
virtual configuration items and their
dependencies
Deploys based on standard baseline host
profiles and proactively monitors and
remediates non compliant hosts
Improve management, tracking and
control over desired state of IT assets
Visibility Impact
Analysis
Policy based
control &
automation
Simplified
Configuration
Adaptive, Actionable
Intelligence Engine
Apply Profile, Monitor & Remediate
External
CIs
Cmdb (CIs)
2009
Infrastructure Management
47. 51
Administer and control large
environments easily
Provide custom workflows for
complex environments
Workflow engine for orchestrating
virtualization
Automate manual, repeatable steps
by drag and drop interface without
scripting
Centralize workflow management for
all processes associated with the
environment
Unified
Operations
2009
Infrastructure Management
48. 52
Enables scale out of managed
infrastructure
Simplifies configuration of new
management servers
Inventory and configuration data
available globally via LDAP backbone
Standard VI Client can access
inventory across entire environment
Roles and Licenses are replicated
across all VC Servers
VI Client supports operations that
span multiple VC Servers
Unified
Operations
Replicated Inventory
and Configuration Data
ESXi ESXi ESXi ESXi ESXi ESXi ESXi
VirtualCenter
Server
VirtualCenter
Server
VirtualCenter
Server
2009
Infrastructure Management
49. 53
Reports on
usage
associated
costs
Provides
insight into
resources you
are paying for
but not using
Identifies
unused VMs
based on
capacity
usage
Verifies
sufficient
capacity
Runs what-if
scenarios to
predict potential
capacity impact
Considers
capacity
forecasts &
predictions
as incoming
VM requests
are received
End-to-End Value Across vCenter
Solutions
Provision DecommissionDeploy Track
Lifecycle Manager
Task
Management
Infrastructure Management
vCenter
CapacityIQ
vCenter
ChargebackvCenter CapacityIQ
51. 55
Development & Testing
Resource Pool
Pre-Production
Resource Pool
Production
Resource Pool
VMware Infrastructure
Streamline Application Dev & Deployment with
Lab Manager and Stage Manager
Provision and transition exact replicas of complex services
throughout the lifecycle with much less effort and hardware
Empower users but keep central control of policy and quotas
Test Integration Staging ProductionDev
52. 56
Policies (SLA)
Quality of Service
99.9% Uptime
Fault Protection
Encryption Enabled
55
Users
2
Servers
1
Database
75
Users
4
Servers
Monitor and Control QoS with vCenter AppSpeed
Monitor application performance
from end-user perspective
Collect and correlate across tiers
and in VI
Automated SLA management
Enables proactive detection of
end user performance issues
Integration with VI enables root
cause and remediation
53. 57
VMware Ready Management Solutions
VMware Ready Management & vCenter
Application Management
Infrastructure Management
vCenter
vCenter Extensibility
vCenter – foundation for
Application Management
& Infrastructure
Management
vCenter Extensibility –
APIs and UI plug-in
architecture to enable 3rd
party integration
VMware Ready
Management Solutions
• Integrated using
vCenter Extensiblity
• Based on VMware
guidelines
• Interoperability
Qualification
Program
55. 59
virtual datacenter OS from VMware
Off-premise
Cloud
vCenter
On-premise Infrastructure
SaaSLinux GridWindows J2EE.Net
VMware Infrastructure -> virtual datacenter OS
Application
vServices
Scalability
Infrastructure
vServices
SecurityAvailability
vNetworkvStoragevCompute
Cloud
vServices
…….
Web 2.0
56. 60
Integrated Provisioning Process
Example
VMware Management
(VI and LCM)Initiate Deploy
Add to
Cluster Verify
Service
Desk
Request 3rd Party
Approval RunBook
Provision
Storage,
NW, etc.
Verify/
Notify
VMware announced a comprehensive roadmap to expand VMware Infrastructure into a virtual datacenter OS (VDC-OS).
The VDC-OS abstract and aggregates all types of hardware of a datacenter – server, storage and network – into a single, shared resource or “internal cloud” that can be allocated to applications that need it most with great precision. The VDC-OS is the ‘production line” for the datacenter in terms of automation and efficiency. IT can run their infrastructure like a just-in-time factory.
It also federates with external clouds of computing capacity, so that any enterprise can run their IT as efficiently as Google or Amazon– at a very low cost with lights-out automation. The VDC-OS is cloud computing for the enterprise made easy.
It has built in application services – for availability, security and scalability that any application running on the VDC-OS can leverage.
VDC-OS is a vital necessity for the modern datacenter. VDC-OS is to the entire datacenter what Windows and Linux are to a single server.
The VDC-OS enables a dramatically simpler computing model that changes the lives of IT people – “from swinging the hammer to pushing a button”.
While the competition is still trying to replicate capabilities that VMware has been shipping for years (such as VMotion), VMware is releasing even more impactful capabilities like VMware Fault Tolerance – zero downtime, zero data loss failover. It is like the competition is trying to bring to market color television, while VMware is releasing HDTV.
The traditional IT stack with tight coupling of software and hardware falls short of customers’ needs to support increasing rate of business change, non-negotiable requirements for 24X7 business resiliency as well as inexorable pressure to reduce cost. Traditional x86 operating systems are tightly bound to discrete single servers resulting in;
- Rigid, inflexible infrastructure that cannot easily respond to changes. Because scaling the infrastructure can take months - for example provisioning a new server often takes 6-8 weeks – companies have reacted by massively over-provisioning, and exacerbating server sprawl.
Availability, security and QoS for applications is a complex task, custom to every application because of the interdependencies of application, OS and hardware
Server sprawl and gross resource underutilization – typical server is utilized only 5% of the time. A negative side effect of server sprawl has been the energy crisis in the datacenter
- Very complex and brittle management model where even routine management tasks take months or incommensurate amount of effort. For example performing regular maintenance on a server or applying a routine OS patch requires taking applications down.
In addition to the challenges with the existing computing model, some important trends create new opportunities and threats for customers:
x86 hardware becoming more prevalent – multi-core processors, very large systems with hundreds of processors, lots of memory etc - x86 hardware becoming more capable than ever all at a decreasing real cost. As a result, x86 systems can support bigger and bigger applications and the share, importance and relevance of x86 systems is on the rise.
Innovation and experimentation with application architectures/stacks. SOA, web 2.0 – multi element, mixed OS applications are becoming more prevalent. There is a lot of experimentation and vibrant innovation in application stacks which makes it hard to predict what that prevalent app architecture in the future is going to be, if there is one. this poses a real challenge for customers to future-proof their infrastructure investment.
Massive cloud datacenters being built by SaaS, cloud, traditional hosting providers such as Google or Amazon. These massive datacenters can provide capacity on-demand very cheaply. However, today to take advantage of this cheap capacity, enterprise applications have to undergo a great deal of customization/porting. Also, customers have to make the difficult choice of either doing things the old way – or completely recreating applications to run off-premise.
<This slide is part 1 of a 3 slide build. It has animations>
What customers really want is to be able to provision an application when needed and enable service levels for that application at the click of a button. They want to set the availability parameter, the security parameter, the performance scaling parameter and then provision it to the place that has the lowest TCO. If that’s the on premise data center – then the app gets provisioned on-premise. However, if the off premise datacenter or cloud has a more available capacity, or better economics, then it is simple to just move the app over there…along with the same service level parameters as before, or maybe with changed ones.
What customers really want is to be able to provision an application when needed and enable service levels for that application at the click of a button. They want to set the availability parameter, the security parameter, the performance scaling parameter and then provision it to the place that has the lowest TCO. If that’s the on premise data center – then the app gets provisioned on-premise. However, if the off premise datacenter or cloud has a more available capacity, or better economics, then it is simple to just move the app over there…along with the same service level parameters as before, or maybe with changed ones.
<This slide is a continuation of slide 5’s animation>
What customers really want is to be able to provision an application when needed and enable service levels for that application at the click of a button. They want to set the availability parameter, the security parameter, the performance scaling parameter and then provision it to the place that has the lowest TCO. If that’s the on premise data center – then the app gets provisioned on-premise. However, if the off premise datacenter or cloud has a more available capacity, or better economics, then it is simple to just move the app over there…along with the same service level parameters as before, or maybe with changed ones.
The technology category that delivers the virtual datacenter is the virtual datacenter OS.
This is the model that delivers the simplicity that IT admins are looking for.
virtual datacenter OS is a reliable, extensible, and manageable software platform that delivers the elastic, self-healing and self-managing datacenter.
The VDC-OS seamlessly aggregates on-premise servers, storage and network into “an internal cloud” – an elastic, shared, self- managing and self-healing utility that also federates with external clouds of computing capacity to free IT from the constraints of static hardware-mapped applications. The VDC-OS guarantees the right levels of availability, security and scalability to all applications independent of hardware and location.
The virtual datacenter OS is an OS because it provides the two basic functions of an OS:
Management of the underlying hardware
Services such as availability, security, performance guarantees to applications
However, the virtual datacenter OS had important advantages over the legacy general purpose OSes that make it the right place to run the applications of today and tomorrow:
Support for all existing and future application stacks without requiring code changes. Existing applications can run as-is on the virtual datacenter OS; new applications can run in the runtime containers defined by developers or directly on the virtual datacenter OS. Given the vibrant innovation in application stacks, the virtual datacenter OS represents the best approach for companies that legacy and future applications can be run on the same platform.
Hardware & location independence. Unlike legacy OSes that are installed on a single server, the virtual datacenter OS is a distributed OS that aggregates pools of industry standard hardware and presents is as a single computer to application. The distributed nature of the virtual datacenter OS makes it more resilient and scalable than general purpose OSes. The virtual datacenter OS also makes applications completely independent of the location where they run.
All the core application services such as security, availability, performance needed by an application are built in the virtual datacenter OS and are enabled for all appplications in a uniform manner without the need for complex customization.
The virtual datacenter OS consists of
A set of infrastructure vServices to seamlessly aggregate on-premise servers, storage and network
A set of cloud vServices to federate the on-premise infrastructure with third party cloud infrastructure
A set of application vServices to guarantee the right levels of availability, security and scalability to all applications independent of hardware and location.
A set of management vServices that allow to proactively manage the virtual datacenter OS and the applications running on it.
Unlike a traditional OS, which is optimized for a single server, the virtual datacenter OS serves as the OS for the entire datacenter. Datacenters of cloud providers and SaaS companies - all on cheap commodity hardware.
The VDCOS enables the dramatically simpler model of computing that our customers have been waiting for. In this new model customers define the desired outcomes, and the computing infrastructure can guarantee these outcomes precisely, at the lowest TCO possible, leveraging either the internal cloud or third party cloud computing services. For example, in order to deploy a new application, any application, an IT professionals would specify application service levels such as response time, security protection, and availability level– and the infrastructure should deliver and ensure these service levels, at the lowest possible cost with low maintenance effort required.
The VDCOS delivers this simplicity.
VDCOS is a software platform that seamlessly aggregates on-premise computing assets into pools of capacity and federates with third party cloud infrastructure to deliver capacity on demand and guarantee the right levels of availability, security and performance scaling to all applications independent of hardware and location.
The VDCOS is an OS because it provides the two basic functions of an OS:
Management of the underlying hardware
Services such as availability, security, performance guarantees to applications
However, the VDCOS had important advantages over the legacy general purpose OSes that make it the right place to run the applications of today and tomorrow:
Support for all existing and future application stacks without requiring code changes. Existing applications can run as-is on the VDCOS; new applications can run in the runtime containers defined by developers or directly on the VDCOS. Given the vibrant innovation in application stacks, the VDCOS represents the best approach for companies that legacy and future applications can be run on the same platform.
Hardware & location independence. Unlike legacy OSes that are installed on a single server, the VDCOS is a distributed OS that aggregates pools of industry standard hardware and presents is as a single computer to application. The distributed nature of the VDCOS makes it more resilient and scalable than general purpose OSes. The VDCOS also makes applications completely independent of the location where they run.
All the core application services such as security, availability, performance needed by an application are built in the VDCOS and are enabled for all appplications in a uniform manner without the need for complex customization.
How did we get here?
VMware started many years ago with simple abstraction of a single server to run multiple virtual machines on the server. We quickly expanded these capabilities to not only abstract away from all hardware elements – server, storage and network but also to unify these hardware elements and present a shared platform for applications. Along the way, we created the technology that liberates the applications from the underlying hardware and also, through this shared, reliable platform delivered built in resilience to planned and unplanned downtime, assured performance guarantees for applications and a highly secure platform for applications to run on. As we build out these services further , we essentially deliver the vision for a shared dynamic utility IT environment that seamlessly federates on-premise and third party cloud infrastructure, and guarantees the right levels of availability, security and performance to all applications independent of hardware and location.
With the addition of new vServices, the VMware platform is becoming a true OS for the datacenter, or virtual datacenter OS. Unlike a traditional OS, which is optimized for a single server and support only those applications written to its interfaces, the VDC-OS serves as the OS for the entire datacenter and supports the full diversity of any application written to any OS, from legacy Windows applications to modern distributed applications that run in mixed operating system environments.
VMware is the best positioned to deliver this technology category.
VMware customers are already seeing the transformative benefits of VMware Infrastructure- the dynamic infrastructure created by VMware saves money, time and management effort on an ongoing basis. Built in availability and load balancing make it easier to manage infrastructure.
As a result, more than 120,000 customers of all sizes around the globe have been able to tackle real IT problems to reduce capital and operating expense, ensure business continuity and become more agile and responsive than ever before.
This is just a sampling of some of the results of the capabilities we provide to our customers.
Customers save up to 60% in total cost of infrastructure on a per application basis – and increase productivity of admins almost 2-3x.
The net effect that we typically see among our customers is that they can meet business requirements at a fairly low cost, and because they can spend a lot less time fighting fires on a daily basis – they are able to manage many more workloads than they would in a physical environment. They are also able to spend more time architecting their infrastructure better for higher levels of productivity.
Supporting data:
. Before VMware, IDC claims an average of about 30 servers per admin. After virtualization, they typically see the number increase 3x.
In the TAM program, we considered collecting this statistic. One company with a very good VM to Admin ratio is Fidelity Investments. They had approximately 300 : 1 ratio (Approximately 1500 VMs to 5 Admins). Bank of America has approx 2000 VMs and between 8 to 10 VI administrators (200:1). The ratios vary widely based on the role of the administrator and the length of their deployment. This is merely anecdotal information and not in depth analysis of our larger customer base, but even IDC has claimed that average server to admin ratios with VMware are 3x or 90 server per admin (John Humphreys, Vforum preso / Virtualization 2.0).
Application vServices – to recap, are services provided by our platform to applications – uniformly, enabled by simple point and click in most cases.
Let’s review the current application vServices provided by VMware and new vServices in 2009
This is a quick summary of everything I am going to cover in this section. Current capabilities are in green and new capabilities are in orange.
Today VMware provides a variety of solutions that shield applications from infrastructure downtime. VMotion protects applications from planned server downtime, HA provides the first line of defense against unplanned server downtime.
Storage VMotion protects applications against planned storage downtime, while Consolidated backup provides a framework to protect against data corruption or data loss
At the interconnect layer, NIC & HBA teaming provide resilience to unplanned component failures
At the virtual machine level, VM failure monitoring provides the automated restart in the case of virtual machine failures
Beyond individual sets of servers/storage, if the entire set goes down, Site Recovery Manager provides the orchestration of recovery from downtime and can be used for planned site downtime/migration as well.
In 2009, we introduce two new solutions that take downtime management with VMware to a whole new level.
VMware Fault Tolerance creates virtual machine “pairs” that run in lock step - essentially mirroring the execution state of a VM. To the external world they appear as one instance (one IP address, one application) – but they are fully redundant instances.
In the event of an unexpected hardware failure that causes the active, primary VM to fail – a secondary, formerly passive VM immediately picks up where the primary left off, and continues to run, uninterrupted, and without any loss of network connections or transactions.
This technology will also work across any application & any OS without modifications, without scripting, and provides a much more cost-effective way of running mission critical workloads than fault-tolerant hardware dedicated entirely to individual applications.
At VMware, we say that “virtual is better than physical” very frequently, and our advanced development work on providing continuous availability for VMs – to enable VMs to keep executing, completely uninterrupted by unexpected hardware failures - is one of the great examples of why this is true. We have taken technology implemented with very complex custom hardware by companies such as Tandem, Stratus etc and delivered it for commodity x86 hardware.
We believe this technology, and all of the other business continuity benefits of virtualization will drive more and more mission critical workloads into virtual machines because they can enable HIGHER levels of availability, at a fraction of the cost & complexity of physical solutions.
DETAILS ( use only as needed)
Limitations of FT initially: dependent on shared storage, Uniprocessor VMs only, mirroring of VMs limited to 2-nodes. Additional overhead also associated with this type of solution. Overall performance impact still TBD, but you can expect more CPU & memory resources will be required to run the 2nd VM, and applications may experience small amounts of added latency.
In spite of initial limitations, longer term trends are in our favor: FT will take advantage of hardware assisted virtualization in CPUs, more and more CPU cores becoming available to offload overhead, and high-speed network improvements like 10gigE to reduce latencies…
Workloads that were protected with automated restart against hardware failures were similarly less than 10% of all workloads because of the cost and complexity of clustering. With HA, we extended automated restarts to ALL workloads in the datacenter , making first level failure protection very very easy and accessible. The workloads that are protected CONTINUOUSLY today against hardware failures are a very small fraction of all workloads, because the cost and complexity of solutions that provide the continuous levels of availability is very high.
We expect that with FT, we will have a similar effect where many critical workloads that are only protected by HA or not protected against physical failures at all are able to gain FT protection.
vCenter Data Recovery is a new tool for VMware Infrastructure users that provides complete data protection for your virtual machines.
- It’s a disk-based solution that’s easy to use and fast to backup and restore.
- It’s built on the VCB API and is fully integrated with vCenter management to enable centralized and efficient scheduling of backup jobs using an intuitive workflow.
Step 1: Backing up your virtual machines is a snap –
- First you schedule your backups directly through the vCenter interface.
<FIRST BUILD CLICK>
- Snapshots are taken at your scheduled intervals and written to near-line storage (local or shared).
- After the first full VM backup, subsequent backups are incremental to save time and disk space.
- vCenter Data Recovery also uses disk de-duplication to further reduce the disk space required for backup storage.
<SECOND BUILD CLICK>
Step 2: The toughest part of the backup process is recovering your apps and data. vCenter Data Recovery makes this process quick and easy –
- When a VM goes down, you go to vCenter and quickly access the directory of backup VMs and files.
<THIRD BUILD CLICK>
- Select which ones you want to recover (individual files or complete VM) and restore in seconds/minutes
- Your applications and data are back up and running!
One important aspect of security is the size of the virtualization layer footprint. ESXi with its 32MB of code, locked down interfaces and independence from a general purpose operating system already provides the most secure way to virtualize.
VMware VMSafe announced earlier in 2008, is a set of APIs that enable protection of VMs by a protection engine that :
Works with the hypervisor to inspect a VM’s mem, cpu and storage from a higher privilege point
Is isolated from the malware
Covers all aspects of security – not limited to network or host.
MORE DETAIL
Security solutions have an inherent problem. Protection engines are running in the same context as the malware they are protecting against and as a result, malware is able to subvert these engines by simply using the same hooks into the system as the protection engine. Worse, with Longhorn and Vista, Microsoft has enabled Patchguard, effectively eliminating the kernel hooks available to both the security solutions and the malware. While this helps, it doesn’t change the fact that malware and rootkits still exist and can run in those environments. The context that these security solutions need to protect against is also not limited to one set of interactions (e.g. attacks from the network and from spyware and from rootkits). Even those solutions that are in a safe context (outside the OS), they can’t see information from other contexts (e.g. network protection has no host visibility).
Security API’s built into the hypervisor allow for 2 key advantages:
Better Context – Provide protection from outside the OS, from a trusted context
New Capabilities – now they can view all interactions and contexts
Now, new security solutions can be developed and integrated within the VMware virtual infrastructure and we can protect the VM by inspection of virtual components (CPU, Memory, Network and Storage). Provides complete integration with VMotion, Storage VMotion, HA, etc. for any new security solution using the API’s. The end-result is an unprecedented level of security for VMs that’s better than the physical infrastructure. These API’s are already being made available to the security ISVs ecosystem.
We utilize VC for role-based privileges to assign protection to any single VM and VMware certifies the solutions developed by our partners to ensure the security VM is created by a real security ISV and not a malicious hacker.
Some potential use cases:
An AV virtual appliance that intercepts all storage IO and is able to scan files as they are read/written from disk. This can be done without loading an AV agent on each machine.
Inline Network Security for each ESX host. Now you can ensure that ALL network IO traffic is inspected by an inline appliance, regardless of your virtual networking setup. This includes even inter-VM traffic and allows state to be transferred from host to host during VMotion so that the security protection is never lost.
Partner solutions that utilize VMsafe have the following advantages over existing security mechanisms:
They can protect VMs without needing to install agents inside each VM
They can perform a multitude of tasks that protect a VM holistically ( monitor VM components on the host, nw traffic through the distributed switch) AV, Firewall, IPS can all be in one appliance.
They can do all these tasks efficiently in a dynamic environment where the virtual machine moves from host to host. For e.g. if a VM is being monitored on a particular host, it carries its security state with it to the next host where another appliance starts monitoring it.
Application scalability is ensured today through DRS shares and reservations, which allow applications to grow and consume as many resources as they needed according to their priority. In 2009 , application scalability is assured without any disruption or downtime with “hot add” . Also applications themselves can scale to 8 virtual CPUs and 256 GB of RAM. What this ensures is that the most intense applications can be virtualized, without being worried about future scalability.
Think of the physical world today – where scaling an app means a complex task involving detailed sizing, procuring hardware, application downtime, then moving the application over to new hardware etc.
Virtualization already made this process easier with hardware independent movement of apps – but with hot add, now applications can be provisioned in a “future proof” manner. As apps grow, as they get more and more intense over time and need more compute, memory or network/storage resources, admins can now scale them up dynamically – no disruption, no complex porting , on the fly.
Infrastructure vServices use industry standard server , storage and network components to create a unified efficient and shared platform. Not only do they abstract away from the underlying hardware, they also aggregate this hardware and present a unified set of resources, which can them be logically carved up and provided to applications in a dynamic fashion, based on business requirements and priorities.
Now we move into what’s new – what are we announcing at VMworld. What we are announcing falls into four major categories, organized by benefits.
The first of these is “VMware’s Next Generation Virtual Datacenter” maximizes infrastructure efficiency to get the lowest TCO of any virtualization solution.
We virtualized server, storage and network to transform it into an efficient and shared infrastructure.
New product capabilities from VMware in 2009 include:
The use of VMDirectPath i/o and network offload to reduce CPU overhead associated with i/o processing. VMDirectPath allows association of a VM directly with the underlying i/o device – so if there are VMs that require constant, uninterrupted access to the device, their path is more efficient
Today we achieve 100k iiops comfortably meeting the requirement of 90% of applications. With paravirtualized SCSI, we optimize storage access and can achieve greater than 200k iops
For such intense applications, we enable scaling to 256GB mem and 8 way vSMP
ESX/ESXi scale to 512 GB of memory and 64 cores enabling very very large consolidation rations
-VStorage is the set of VMware and partner capabilities that enable the most efficient utilization of storage in virtual environments
This is a quick summary of everything covered in the infrastructure vServices section.
At the highest level – new vCompute vServices further reduce overhead and deliver the most efficient way to virtualize.
New storage vServices deliver the most efficient way to use storage in virtual environments
New network vServices deliver the most optimal way to integrate networking in virtual environments
Several mechanisms already exist in VMware ESX making it the most efficient way to virtualize. ESX and ESXi virtualize every resource with the lowest possible overhead
A purpose built scheduler and use of hardware assist technologies for optimizing CPU access
Transparent page sharing/ballooning for the most effective usage of memory and hardware memory management assists for the lowest overhead
In 1H2009, we introduce several new techniques that supplement existing mechanisms for the highest throughput for i/o
With 3.5, we already introduced offload technologies that offload TCP processing lowering CPU overhead associated with i/o
We now introduce VMDirect technology that allows a VM direct access to underlying devices – for the workloads that need constant access to the underlying hardware, this enables low CPU overhead by allowing the VM direct access to the device.
Para – virtualized SCSI devices optimize access to storage for the workloads that require very frequent access to storage – by making the virtual storage devices virtualization aware, and thereby reducing access times.
ESX scales now to support 128 cores and 512GB of physical RAM – so you can virtualize very large scaled out servers and run them with the most efficiency
Virtual machines themselves can now scale to be very large – we covered this before.
Some of the results of the inherent efficiency of VMware’s virtualization are visible today:
When you run a web server on a physical server and scale it up from 1-6 CPUs, the performance levels off because the physical app is not able to scale very efficiently and utilize all the processors well.
When you run multiple web servers ( 1 CPU per VM) on the same physical server virtualized, the output increases linearly with the number of web servers – this is because our purpose built scheduler imposes a very low overhead, scales linearly and is able to utilize multiple cores much more efficiently.
So not only do you not see any dropoff in output, you actually get better performance than physical!
In terms of absolute numbers, with 3.5, we address the performance requirements of a large section of workloads today. In 2009, we now allow the small fraction of workloads that are intense enough to require upto 200,000 iops or 8 virtual CPUs.
Peak means peak hours: 7 am – 7 PM
Virtualizing and pooling all elements of the infrastructure.
In addition to the most efficient single server virtualization, VMware optimizes the efficiency with which IT infrastructure is shared across many workloads.
Today we do this through DRS which optimizes virtual machine placement based on CPU and memory requirements.
DPM optimizes for power consumption by consolidating workloads and turning servers off when virtual machines don’t need as many resources. ( production level support for DPM will be in 1H2009)
In 1H2009, we introduce the vNetwork Distributed Switch that aggregates networking at a datacenter level and enables datacenter wide networking services.
And we introduce vStorage – our umbrella set of storage technologies that optimize the use and management of storage in virtual environments.
vNetwork is a collection of networking technologies VMDirectPath, NetQueue, and Distributed Switch for optimally integrating networking and I/O functionality into VMware Infrastructure.
vNetwork Distributed Switch enables the network to be treated as an aggregated resource… much like what we do with compute and storage already today
DVS moves away from single host virtual switch management, and moves management up to the datacenter level. With DVS, no longer will you be configuring virtual switches on every host. Instead, with DVS, you’ll be managing a single global entity, with a single namespace and globally assured configuration, and you’ll be attaching VM’s only to that cluster or datacenter-wide entity. DVS is a new type of virtual switch which spans the entire Virtual Infrastructure…analogous in many ways to a “stacked switch”.
Today, when virtual machines migrate from one host to another,
Network statistics don’t follow the VM after it migrates
Value-added services like inline filtering, failover teaming, etc. may not follow the VM
By creating DVS, a single cluster-wide global virtual switch, we lay the foundation for a new generation of properly mobile, networking services.
DVS lays a virtual networking foundation for both VMware and partners to build richer, mobility-friendly network services in which policies, rules, value-adds, metrics and statistics become mobile with the VM.
VMware is enabling third party partners to create virtual switches; in actual fact, they will be able to create distributed virtual switches.
Cisco , for example , is announcing a product to be delivered in 2009 which uses our vNetwork Distributed Switch framework to plug into VMware virtual environments. Through this new virtual switch, Cisco will be able to extend the same qos assurance , security assurance that they provide in physical environments to virtual environments at a VM granularity.
Network admins will now gain visibility and control into VMware environments, and server admins can be assured of networking services on a per VM basis.
vStorage is the technologies and interfaces that VMware provides to enable simplified management of the storage infrastructure that customers use in the VMware environment. vStorage integrates and optimizes storage infrastructure by delivering a combination of VMware technologies and storage partner capabilities.
vStorage includes the following VMware technologies that allow customers to simplify and optimize management of the storage used by virtual machines:
Currently available technologies including VMFS and Storage VMotion
Future technologies including Thin Provisioning, Linked Clones, VMDirectPath for storage, and enhanced management functionality in VirtualCenter.
vStorage also includes important API’s and interfaces that enable VMware and storage partners to integrate and optimize their storage partner technologies with the VMware platform:
Storage Virtual Appliances extend the capabilities of the VMware Datacenter OS with technologies from storage partners in the form of VMware Ready virtual appliances.
vStorage API’s make it easy for customers to directly leverage the advanced capabilities of their storage infrastructure with virtual machines. By leveraging these API’s, partners can deliver seamless integration of their tools and technologies with the virtual environment.
vStorage Thin Provisioning
Another important new capability that maximizes storage utilization by virtual machines is the vStorage linked clone technology. This technology will reduce duplicate storage of virtual machines’ data, transparent to the virtual machine, by allowing multiple virtual machines to share common data in a single base disk while maintaining separate storage for the data written by each virtual machine. This technology will also enable simplified patching—patches that affect just the base disk can be applied once to the base disk and seen by each virtual machine without needing to be applied individually to each virtual machine.
This technology will first be leveraged by the new VMware View Composer product to significantly reduce storage costs for VDI deployments, allowing similar virtual machines to share a common base configuration disk while still maintaining their individual configurations and data.
Storage virtual appliances deliver key storage functions from storage partners in the form of virtual appliances. Storage Virtual Appliances make it easy for small environments—such as SMB, remote office, and branch office deployments—to obtain the benefits of technologies that could include shared storage, data protection, data deduplication and others that were previously not possible for them to deploy. They accomplish this by simplifying deployment and configuration through pre-configured virtual appliances and by eliminating the need for external hardware for this functionality. As these deployments grow, customers have an easy migration to partners’ physical hardware that provides the same capabilities for larger and more demanding environments.
The VMware Ready program provides validation for key types of storage virtual appliances so that customers can be assured that they can confidently deploy storage hardware and storage virtual appliances that have been tested and validated for the VMware Infrastructure platform.
vStorage API’s are a set of new and existing API’s that enable storage ecosystem partners to integrate and optimize their products with VMware Infrastructure. These API’s enable storage vendor’s products to understand and optimize for the virtual environment. They simplify and automate the manual handoffs and processes between the virtual environment and the storage infrastructure to deliver simple, coordinated management of virtual infrastructure and storage infrastructure.
The vStorage API’s include the following:
vStorage API’s for array integrations are new integration API’s for VMware Infrastructure that will provide customers the ability to fully leverage the advanced capabilities of their storage systems with virtual machines at a far more granular level than before. These integrations will enable customers to leverage array-based capabilities, such as snapshots, provisioning, replication and restore with individual virtual machines in conjunction with the clustering and pooling capabilities of VMware Infrastructure.
vStorage API’s for multipathing will provide customers the ability to integrate advanced load balancing capabilities provided by leading partners’ multi-pathing software with their virtual environment.
vStorage API’s for Site Recovery Manager currently provide customers with tight integration between VMware Site Recovery Manager and array-based replication, enabling customers to define a custom recovery process and automate the restart of their virtual machines after disasters without having to worry about correctly coordinating the virtual environment failover with storage replication failover.
vStorage API’s for Consolidated Backup enable customers to automate and streamline non-disruptive virtual machine backups using standard backup software products.
vStorage API’s for storage management include current and future interfaces to VMware Infrastructure that enable integrated management of storage hardware platforms and the virtual environment through management tools from VMware and storage partners, eliminating the complex coordination and manual processes previously required to manage storage within a virtualized environment. Examples include integration of storage management tools via VirtualCenter plug-ins and VI SDK interfaces that storage management tools use to understand the use of storage by the virtual environment.
Cloud Computing is the next disruptive trend in how IT will consume and manage their infrastructure and applications. As the industry continues to morph, vendors, analysts and press will arrive at their own versions of what cloud computing is and how it will affect IT professionals. In talking with many of our customers, we feel that cloud computing is the ability to:
acquire elastic computing capacity on demand in a pay per usage model.
Although there are cloud computing providers out there today, IT organizations still have challenges with how these providers will handle critical production level applications- and have been slow to adopt it for those uses. Instead these clouds are populated by developers and startups looking for quick and cheap infrastructure. In order for enterprises and SMBs to adopt the cloud for production level uses, these limitations need to be addressed. These limitations include
the inability to move an application from a provider in the event of something like poor SLAs, or to simply move applications between on and off premise to accommodate things like Disaster Recovery.
The inability to provision capacity or services on demand
And the incompatibility of applications that need to be written for a cloud environment, vs what would be used in-house.
VMware vCloud, a set of cloud computing platform technologies for enterprises and service providers with broad support for existing and new applications. The vCloud initiative, backed by more than 100 service providers including BT, Savvis, Sungard, T-Systems, and Verizon, helps businesses move to the cloud how they want, when they want, and as much as they want—without disruption.
There are 4 core pillars to the vCloud:
Most Efficient and Flexible Technology Platform for Cloud Computing
A highly efficient, flexible and automated infrastructure is critical to enabling cloud computing both in the enterprise and on the public internet. vCloud services leverage unique capabilities of VMware Infrastructure to enable optimal delivery of cloud computing:
Best in class hypervisor that provides highest single server efficiency:
The ability to aggregate cost-effective x86 servers, storage and network into a single platform to achieve the highest cluster efficiency:
Highly scalable infrastructure management via fully automated datacenter: Utility, pay-for-what-you-use infrastructure:
Broad Application Support
vCloud services give enterprises choice and flexibility by enabling any application from legacy Windows NT to modern day Ruby-on-Rails to be deployed on-premise or off-premise—without disruption. Unlike other compute clouds that require applications to be built specifically to a single cloud computing platform and require complete rewrites of existing applications, millions of existing applications currently running on VMware Infrastructure can run on vCloud services without modification. For any size of organization, this application compatibility provides the flexibility to run applications where it makes the most sense, on premise or off with vCloud.
Industry Leading Service Providers
vCloud enables all kinds of customers and use cases to leverage cloud services by enabling an extensive and diverse ecosystem of services. With over 100 partners worldwide, the VMware Service Provider Program (VSPP) extends the vCloud platform and creates mobility and interoperability across common services built by VMware, service providers and customers themselves.
vCloud Technologies
vCloud technologies will connect internal datacenters and external service provider offerings, enabling enterprises to move between on-premise and cloud-based services, to scale the infrastructure needed for peak load, service level management, and disaster recovery without expanding datacenter capacity.
vApp is the UPC barcode that specifies policies for applications running on the VDC OS
Anyone using VMware Infrastructure can use the vApp to encapsulate a multi-virtual machine application
ISVs can use VMware Studio to create vApps that can be automatically updated and maintained by VMware Infrastructure
vApp turns new and existing application into self-describing and self-managing entities. vApp leverages OVF, an open industry standard, to specify and encapsulate all components of a multi-tier application as well the operational the policies and service levels associated with it. Just like the UPC bar code contains all information about a product, the vApp gives application owners a standard way to describe operational policies for an application which the VDC-OS can automatically interpret and execute.
The dynamic fluid environment created by the virtual datacenter OS requires management capabilities that are:
Aware that resources are inherently shared.
Aware of the mobility of applications within and across locations.
Aware of the service level contracts between apps and the infrastructure.
vCenter products and new capabilities to be introduced in 2009 equips companies with a comprehensive set of infrastructure management capabilities to automatically provision new VMs and vApps, to ensure compliance with established configuration standard, right-size every element in the environment, and allocate the costs back to the business.
Simplified
Managing IT services independent of their infrastructure
Automated
Accelerating processes while ensuring compliance
Responsive
Re-configuring quickly to guarantee application QoS
Resilient
Recovering from any failure without disruption
The Role of VirtualCenter
VI is a distributed architecture with VirtualCenter as the cluster master
Provides resource management and workload mobility across the virtualized environment to aggregate resources into a consolidated shared infrastructure
VC represents an abstraction layer for acting on and managing this shared infrastructure
VC is the foundation for VMware’s other management solutions providing:
enterprise readiness foundation (scalability, availability etc.)
shared services (workflow engine, event bus, policy mgmt, etc.)
integration (APIs, GUI extensions, etc.)
VC is a partner-extensible platform for managing the virtual infrastructure
Partner management tools can register management capabilities with VC console
Partner tools can use VC as execution engine for VMware environment
vCenter delivers comprehensive infrastructure management capabilities for virtual environments and also provides extensibility interfaces to enable it to plug into broader physical + virtual datacenter management frameworks
VMware vCenter integrates with industry-leading systems management solutions for enterprise-wide system management. What’s new?
The VMware Ready Management solution program provides customers with validated partner solutions that further simplify end-to-end physical and virtual management with assurance of interoperability and consistent support and documentation. This enables customers to more quickly deploy management solutions and drive new levels of agility and automation in the datacenter.
For 2009, we introduce broad ranging infrastructure and application management capabilities that simplify virtual infrastructure management
-Application Service Level Management with Bhive Conductor – renamed to vCenter AppSpeed
-New capabilities on Virtual Infrastructure Management include ( see list above)
-End to End Physical and virtual management through seamless integration with leading systems management vendors
Let’s start with vCenter Infrastructure Management:
Today – we automate the provisioning and decommissioning process with vCenter Lifecycle Manager.
We will be adding chargeback functionality into VI that will include the ability to setup an organizational model for grouping VM’s by organizations, ownership, etc. In addition to the organizational model, chargeback functionality will allow charging based on both the allocation of resources or based on actual resource utilization. We plan to include the ability to monitor chargeback based on memory, CPU, disk, disk i/o, and network i/o. In addition, the solution will be able to add in fixed costs to be able to track costs associated with things such as software licenses, real estate, power, etc.
Lifecycle Manager will integrate with the chargeback functionality to make chargeback and fundamental part of the provisioning process. The goal is to allow requesters to understand the costs of VM deployments at the time of requests. This will help requesters make better decisions on what resources to deploy. This will also include the ability to tie dollar costs to different virtual services such as HA, Backup, DR, etc as well as define different cost structure for different levels of service for where VM’s get deployed.
VMware Capacity Manager is a planning, design and decision-support tool for VI capacity. It is a plug-in to VC for ongoing capacity management of your virtualized environments. VMware Capacity Manager provides the following values:
Visibility: how much capacity is being consumed, how much do I have left
Efficiency: align capacity resources, reclaim unused or over-provisioned capacity
Predictability: when will I run out of capacity, what happens if I add/ subtract a host from this cluster?
Features include:
Monitoring: view capacity and utilization in virtual datacenter – how much available, actively used, and which VMs are consuming
Modeling: leverage What-If analysis to simulate cause-and-effect of adding/removing capacity
Trending: Forecast timing of capacity shortfall – when existing resources will run out
Optimizing: Identify and reclaim unused or over-provisioned capacity
Converting: Determine physical systems that are the best candidates for virtualization
Differentiation between VMware Capacity Planner and VMware Capacity Manager:
VMware Capacity Planner is a pre-virtualization tool used in physical environments. VMware Capacity Manager is a post-virtualization tool used for ongoing management of virtualized environments.
VMware will be adding functionality in configuration intelligence with the goal to reduce risk, downtime & cost of configuring and managing the virtual infrastructure by continuous configuration ‘visibility’, ‘automation’ & ‘control’.
We were hearing feedback from customers in 3 key areas : want richer information from VC, want a better way to get it and want to consistently configure entities in VC. Other points of feedback and drivers include:
In order to ensure minimal business impact due to IT changes, customers want greater visibility across their infrastructure stack, including configuration information spanning applications to supporting hardware. Having a single view of configuration data allows them to understand how the infrastructure is configured and track changes over time. This is a critical element to understanding how configurations drift, which can help with identifying potential problems or bad configurations. However, existing IT management tools often require large administrative efforts to provide this view, which can quickly become out of date as configurations across the stack change. As a result, customers end up with outdated views into configurations that often require multiple tools to get a complete picture.
Customers want to ensure a higher level of confidence associated with configuration changes within their infrastructure. They want to know that an infrastructure change (at any level in the stack) will not result in performance problems and outages for critical applications. This requires customers to have impact analysis capabilities that allow them to understand relationships between infrastructure components and how they link together to support business applications. In addition, they need to be able to leverage this to do predictive analysis to understand the upstream and downstream effects of making a configuration change before the change is actually made.
Even with change policies and process in place, customers still face unexpected infrastructure outages as the result of unplanned changes (i.e. a DB administrator updates a simple database script without approval, which brings down a production app). In order to ensure proper policies and procedures are followed, customers want to be able to lockdown configurations so rouge changes cannot occur within the environment. In addition and more importantly, administrators want to codify and automate processes around their operational and configuration best practices to ensure reliability and repeatability.
Customers want increased standardization of their IT assets and configurations to lower administrative costs and errors associated with manual system configuration. Today, many IT systems are configured by groups of administrators that span reporting structures within the broader organization. This can often result in miss step or improper configuration setups that cost our customers money in terms of trouble shooting, diagnostics, and downtime. To remove this challenge, customers want automated system configuration and baselining to ensure systems are consistently configured across their environment. Furthermore, they want tools that can quickly detect out of compliance configurations and the ability to do automated remediation to bring them back in line. This is also important for regulatory initiatives associated with SOX, HIPPA, etc.
Business value for customer: ability to guarantee availability of business critical applications through automating system configurations and controlling how and when infrastructure changes are made. Also being able to determine impact of a configuration change all the way up to the application, business service and being able to prioritize and take necessary actions is important.
At the host level, VMware Host profiles simplifies the configuration of ESX hosts – by allowing the creation of standard “gold” profiles of ESX server, storage, network and security configurations and their application to other hosts. It also allows monitoring of ESX environments for conformance to these profiles and remediation.
VMware Orchestrator is an automation orchestration tool that enables you to put together, via an easy drag & drop interface, automated workflows of tasks and processes specific to your needs and environment. For example, one advanced VMO user created a workflow that enabled datacenter migration in the click of one button. This was a workflow that performed a number of VC operations on each VM in a cluster, including a VMotion, until the entire set of VMs had been migrated to a cluster in another datacenter.
Ability to schedule and automate tasks through building blocks that represent all 800+ VI processes
Some of the feature enhancements that will be delivered in the future include:
Simplified UI option for environments that don’t need complex scripting logic.
better integration with the VC client so in the future, in context right click on an object in the VC client is possible. For example , right clicking on an ESX hosts can show what workflows are applicable to it, and can be run.
The ability to pass in multiple objects into a workflow, making batch workflows possible.
Increase the scalability of the web presentation layer to handle up to 200 simultaneous users, and to be able to scale that horizontally to 8 servers (totaling 1600 web users). Since customers are increasing using VMO to provide VM owners an interface to VC, this allows us to handle expected demand.
So here’s how the various products fit into a end to end workflow: (slide is self explanatory)
Let’s look at application management with vCenter
The overall objective of IT Service Delivery is to get a new or updated application into production (or in the case of an ISV to market) as quickly as possible while minimizing risk. Demands for speed are increasing but accelerating manual processes can result in poor software quality which in turn causes downtime, lost revenue, higher support costs and additional work.
The actual number or names of stages may differ, but typically a company follows a process that includes development, testing, integration (integrating the new patch or code into the overall environment), staging (for performance testing and tuning) and user acceptance testing with end users before rolling it out into production. The traditional approach companies take to move applications through the series of stages or phases shown here is very expensive, manual time consuming and error-prone:
Expensive: A lot of hardware is required. Application development and QA teams use 2-3 times the number of servers required for production. Integration and staging phases require shadow instances of the application because different people work on each stage in parallel. For example, for every server in production for a typical SAP deployment, 7 servers are required in pre-production phases.
Manual and time consuming: Transitioning between pre-production stages requires manually re-building systems. For example, a developer sends a bug report and the tester rebuilds the exact configuration and tries to replicate the bug, often unsuccessfully. Another example is when systems are moved from integration to staging they need to be performance tuned on systems similar to production versions.
Error-prone: Systems experience “configuration drift” from production systems. If a change is implemented on the production instance, it is not always propagated through the rest of the shadow instances. Drift can be caused by process such as patching, or by things that are legitimately different. The other challenge is that of keeping the systems in consistent known state – if multiple users have access to the same systems, it’s not only likely, it’s practically guaranteed that the systems will become “dirty” – settings will be inadvertently changed for example, which means that the next person who uses it doesn’t really know the state of the system.
Today companies deal with the problem by either making changes directly on production systems with their fingers crossed or they deploy very stringent processes and controls that slow things down. Obviously neither approach is optimal.
VMware solves this problem with a suite of products designed to meet the unique needs of developer and QA professionals (Lab Mgr) as well as app and infrastructure administrators (Stage Mgr). I won’t explain the details of each product here but generally they allow customers to provision, share and seamlessly transition exact replicas of complex systems with self-service portals and pools of shared infrastructure that can be easily repurposed. A few examples might illustrate this point:
With a single click in the Lab Manager portal a developer requests a five-tier environment and is working with the application in less than a minute. The developer shares a bug encountered by sending a “livelink” to QA where they troubleshoot using an identical environment.
With a single click in Stage Manager an infrastructure admin “promotes” an entire service configuration from one stage to the next, ensuring consistency of systems. Stage Manager provides a global view of all IT Services, their stages and all associated projects, configurations and resources. The application owner can see at a glance where each application is on the path to production.
In both cases IT maintains centralized control of key aspects such as policies and quotas.
The end results are significantly less hardware, much smoother transitions, better quality software and faster time to market.
Lab Manager Details:
Let’s say that a developer needs a particular five-tier environment – Application, Database, Web, Single-Sign-On servers and a client system. Prior to using Lab Manager, he would submit a trouble ticket to the IT services team requesting that such a system be provisioned for his use, and then he would wait - perhaps a day, perhaps a week – for that request to be fulfilled. In some cases provisioning requests are denied, because there is no more hardware available in the lab or budget to purchase new equipment.
With Lab Manager, you can “check out” a system instantly. As you see here, at the heart of VMware Lab Manager is a shared image library. Each entry in the library is a “configuration” – that is, the complete set of inter-connected virtual machines needed to run a given software system, even if that software system spans multiple servers.
As a Lab Manager user, the developer simply clicks on the library entry that he needs, and in about 30 seconds, the configuration is deployed, enabling him to begin his development or testing duties without delay, and IT admins to spend their time to work on something more meaningful.
As a Lab Administrator you set up the parameters for the lab users, such as access rights, storage quotas or the number of VMs a user can deploy, and within those limits set by IT, lab users can deploy the system configurations they need in self-service mode. You can also impose time limits on users that would free up lab resources if configurations are not used for certain periods of time.
Stage Manager Details:
Before we dive into the capabilities of Stage Manager, let’s take a minute to talk about Stage Manager’s “domain”.
To roll out an IT service or business application into production, you would either develop the application in-house, or customize packed application software. After the coding or customization is done, there is a series of steps before the application can be put into production. The same process can also be applied to systems already in production that you want to patch or upgrade.
The problem of pre-production infrastructure management is solved by:
Consolidating physical pre-production resources and using VMware Infrastructure to set up resource pools to support each stage of the lifecycle
Run IT services and business application software as virtual machines on VMware Infrastructure
Use Stage Manager to manage the transition process of the service configurations
With Stage Manager, you can visualize the service transition process and quickly shift entire service configurations from one stage to the next by a single “promote” operation.
By transition the entire service configuration from one resource pool to the next, Stage Manager ensures consistency, eliminates risk and repetitive provisioning and configuration tasks, because the service configuration remains unchanged as it transitions through the lifecycle. Before Stage Manager, a lot of validation work had to be done to ensure consistency of systems after a service transition.
Stage Manager keeps an audit trail of service operations and configuration changes. The audit trail can be used to ensure compliance with regulatory requirements.
Granular user access roles and rights ensure that service users can interact with only those service configurations in stages they’re authorized to access, and that only service administrators can perform higher level configuration tasks.
When you need to make changes to production systems, Stage Manager allows to take a snapshot of the current service configuration in the archive.
You can then clone a production system, deploy it in a pre-production resource pools, apply your patches and transition the patched system through the release cycle. If the patch is tested successfully, you can either promote the patched system into production, replacing the previous copy, or apply the patches directly to the current production system.
The service configuration clone is an exact replica of the production system, including all network settings, host names and IP addresses. Stage Manager deploys the clone in a “fenced network” to prevent network collisions from happening. The “fenced networking” feature allows simultaneous deployment of multiple instances of the exact same service configuration which allows multiple teams to work in parallel without interrupting or conflicting with one another.
Finally, Stage Manager provides a global view of:
all l IT Services under its management.
the stages associated with each service,
the configurations that exist in support of projects within each stage
The resources associated with each stage and service
It allows the application owner to understand at a glance where each application is on the path to production.
vCenter AppSpeed used Bhive Conductor technology to deliver proactive monitoring and remediation of application performance from the end user perspective.
vCenter AppSpeed collects data across various tiers of an application and tracks it for deviation from SLAs. It is integrated with VMware Infrastructure features such as DRS to help remediation of issues – for eg: it can adjust DRS allocations to a VM, if it appears that the VM performance is suffereing due to lack of a particular resource.
vCenter is an extensible platform with numerous interfaces to enable partners to plug into vCenter environments. In 2008 – VMware Ready Management Solutions will highlight partner integrations with vCenter that interoperate easily and work out of the box with vCenter.
This slide demonstrates the width and variety of the VMware partner ecosystem that have excellent integrations with vCenter APIs
For every vService in VMware, there is a corresponding API - The VMware Ready program is the umbrella initiative that highlights partners that have value added integrations with VMware Ready environments and that provide these integrations themselves.
VMware Ready is a comprehensive initiative providing customers with the assurance that partner solutions have passed VMware-specified integration or interoperability criteria and are technically ready for optimal use with VMware virtual datacenter OS and desktop solutions.
As a result customers reap the benefits of:
Higher productivity and reduced risk based on validated partner offerings and “known-to-work” configurations
The most comprehensive front-to-back solutions with the largest ecosystem of virtual datacenter OS partners
VMware Ready unifies the multiple technology partner programs and designations around a single concept: a logo program for designating to customers the partner technologies and products that are customer ready: technically supported or particularly optimized for use with:
Core VMware virtual datacenter OS (VDC-OS) platform
Servers
I/O and Components
Networking
Storage
Application Security
Cloud Solutions
Virtual Application and Infrastructure Management
This is an example of what an integration with a partner datacenter management solution looks like.