Module 2 App Assessment & Migration Planning Flashcards
First layer of the cloud stack
Applications
Drive the decisions made throughout the rest of the stack in the VDC
Applications
Has the most to gain from a well designed VDC
Application layer
How can application deployment and management be simplified?
By using techniques and technologies available in other layers.
How application uptime can be enhanced
By using features available in the hypervisor
First step in application assessment
To understand the applications in the environment, their value to the organization, and their requirements.
Examples of application requirements
Data security, accessibility, latency, etc.
Four modern application delivery challenges
- Little attention to performance during development
- Security vulnerabilities
- Distributed employees
- Distributed applications & associated complexity
Five new application delivery challenges
- Movement away from monolithic architectures
- Web 2.0 & rich content application requirements
- SLAs
- Legacy apps and application silos
- Consumerization of IT
Three virtualization technology opportunities
- Increase flexibility
- Reduce cost
- Tighten security
True or False:
Generally, the role of application performance & robustness has been pushed down the stack to the compute, network, and storage layers.
True
How can employees be distributed?
Across organization facilities, small field offices, home offices, or at customer sites using mobile devices.
How can applications be distributed?
With multiple presentation layers, residing on separate machines and locations from the business logic layers and data access layers.
Two examples of Web 2.0 systems
How can IT eliminate traditional costs and become leaner?
Through consolidation and standardization
Makes a customer dependent on a vendor for hardware or software, creating a situation where they are unable to use another system without incurring substantial costs or data loss.
Vendor Lock-in
Five items to evaluate to determine if vendor lock-in exists
- Proprietary hardware, software, APIs, file formats
- Software limited to certain hardware platforms or configurations
- Abandonware
- Availability of source code, documentation, skillsets
- Licensing
Three legal issues pertaining to vendor lock-in
- Government laws
- Commercial licenses
- Open Source licenses
True or False: Vendor lock-in is limited to large hardware / software / systems companies.
False.
Three vendor lock-in scenarios for software
- Where software only runs on specific hardware.
- Where software uses unpublished APIs.
- Where software uses undocumented file formats.
True or False: Systems developed in-house can lead to vendor lock-in scenarios.
True
Category of software in which the product is not being maintained due to the vendor stopping work, going out of business, the in-house developer leaving the company, etc.
Abandonware
True or False:
Hardware can also fall into the category of abandonware.
True
Makes it much easier and less costly to work with vendor lock-in situations
Availability of source code
Contributing to the availability of software
Rise in popularity of the Open Source Software (OSS) movement
Included in many Open Source Software licenses
Responsibility to submit any changes made in the code back to the OSS community.
Three areas of savings
- Reduced power, cooling, data center space.
- Better management of licensing costs (be careful of VM sprawl).
- OPEX
VDC Migration Planning Process
- Survey the environment.
- Identify systems that can be combined, commoditized, replaced, or obsoleted.
- Plan goals, using a phased approach.
Question that must be asked when converting to a VDC
How much does it cost to save money?
In other words, what is the up-front CAPEX that will be required to eventually save money?
How should migrating to a VDC be done?
Perform the migration in phases, migrating a few applications and systems at a time.
Five considerations in VDC migration planning
- Determine phases of migration.
- Migrate a few applications and systems at a time.
- Look for overlap and underutilization of resources.
- Ask if some systems might be combined or replaced.
- Identify servers to eliminate or reuse.
Technique that can take advantage of the dynamic nature of the VDC
Combine systems when demand is not high, then scale out as demand increases (since many IT purchases are done in order to handle maximum load).
Five migration plan goals
- Use a phased approach
- Free up hardware and repurpose systems
- Consolidate the network
- Reduce licensing and maintenance costs
- TIme migrations when capacity demand is low
Tricky part of converting the data center into a private cloud
Migrating the legacy, business critical application(s).
Three details associated with Traditional Data Center to VDC phase
Identify applications.
Is decoupling possible?
Rewrite apps?
Detail needed for VDC phase
Application variability and latency
Details associated with VDC to Cloud
Migrating applications to Cloud.
Multi-tenancy.
High availability and eventual consistency (elasticity).
Bursting
True or False:
Many large data centers have the infrastructure to become a Cloud.
True
Becomes even more of a deterrent when an organization is looking to outsource their IT to the Cloud
Migrating the legacy, business critical applications
Three options to consider for application migrations
- Does the cost and effort of application rework, testing and redeployment outweigh the cost of maintaining the legacy app in-house?
- Can the app run in the Cloud infrastructure?
- Once migrated to the Cloud, consider vendor lock-in in case the app needs to move from the Cloud.
Dependencies for successful deployment of Clouds within organizations
- Can the application be decoupled from the physical server (i.e., can the app be virtualized)?
- Can the physical resource meet the demands of the application.
- The application may be legacy and only operate on legacy hardware that is incompatible with the Cloud resources.
Most challenging element for moving the VDC to the Cloud
Applications
Advantage of moving toward the private Cloud
There may be little-to-no redesign efforts required for applications.
Six general application types
- Legacy
- Enterprise
- Big Data
- Web Scale and Horizontal
- Mobile
- Desktop Virtualization
Dated application that continues to serve a business need.
Typically resides on older, slower servers.
Legacy application
Application type that:
- Solves and enterprise-wide problem.
- Software typically used for: order processing, procurement, production scheduling, customer information management, and accounting.
Enterprise Application
Application type that:
- Processes large quantities of data.
- Technologies include: Massively Parallel Processing (MPP) databases, data mining infrastructures (e.g., Apache Hadoop Framework), distributed file systems, distributed DBs, MapReduce algorithms, cloud computing platforms, Internet, & archival storage systems.
Big Data Applications
Application type that:
- Software uses over a broad group of users.
- Examples: word processing, bookkeeping software, social network
Web Scale and Horizontal Applications
Application type that:
- Usually runs on alternative end points such as phones and tablets.
- Uses web technologies: HTML, XML, etc.
Mobile Applications
Application type that:
- Adds a virtualization layer between the server and the desktop.
- Provides greater accessibility, tighter security, and ease of management.
Desktop Virtualization Application (VDI)
Six classifications or categories of applications
Legacy apps Enterprise apps Big Data apps Web Scale and Horizontal apps Mobile apps Desktop Virtualization apps
Applications that typically do not adapt well to a VDC and can take significant work to migrate.
Legacy Applications
What can offset the cost of maintaining legacy apps?
Removing the legacy apps from obsolete hardware.
Disadvantages of Integrating Legacy Apps in the VDC Plan
- Hard to migrate & integrate into the VDC and Cloud.
- Typically built on non-x86 platforms.
- Tightly coupled w/ underlying platform & technologies.
- Application re-work is manual and time intensive.
Advantages of Integrating Legacy Apps in the VDC Plan
- Remove silos of infrastructure dedicated to a single application.
- Move to a modern platform - improve performance & longevity.
- Reduce IT resources needed to manage and maintain legacy apps.
Integration of legacy apps in the VDC plan can be both…
- a source of tremendous gains in a move to VDC.
- a source of great expense in the process.
Three Legacy Application Migration Challenges
- Incompatible dependencies
- Interaction among shared resources - unanticipated problems
- Cross-platform effort
Examples of incompatible dependencies when migrating legacy applications
Libraries
Middleware
Kernel modules
Drivers
Two questions to consider concerning application migrations
- Does the cost & effort of application rework, testing, and redeployment outweigh the cost of maintaining the legacy app in-house?
- Can the application run in the VDC or Cloud infrastructure?
True or False: Considering industry trends, data for legacy applications has little, if any, growth.
True.
Implication: It may not warrant the level of effort & cost a migration may impose.
2.11
5 Enterprise Application Business Characteristics
- Mission-critical (downtime measured in $)
- Typically enterprise wide (impacts all users)
- Dedicated resources to manage & maintain
- Distributed - may span VDCs
- Stringent SLAs & recovery objectives
5 Enterprise Application Technical Characteristics
- Clustered
- Multi-processor
- High memory usage
- High bandwidth requirement
- Large storage footprint
True or False: Traditional horizontal applications, such as email, are increasingly considered mission critical among organizations.
True
EAI
Enterprise Application Integration (EAI)
Established the sharing of functionality and data between separate islands.
Enterprise Application Integration (EAI)
3 Approaches to EAI
- EAI Via Mesh
- Enterprise Service Bus (ESB)
- Service-Oriented Architecture (SOA)
SOA
Service-Oriented Architecture
ESB
Enterprise Service Bus (ESB)
Middleware layer that acts as an intermediary between separate systems
Enterprise Service Bus (ESB)
Architecture where systems are designed as reusable modules that can be called when and where needed.
Service Oriented Architecture (SOA)
Links disparate apps within an org to simplify & automate business processes to the greatest extent possible, while at the same time avoiding the need to make changes to existing apps or data structures.
Enterprise Application Integration (EAI)
Links disparate apps within an org to simplify & automate business processes to the greatest extent possible.
Enterprise Application Integration (EAI)
Avoids the need to make changes to existing apps or data structures.
Enterprise Application Integration (EAI)
Challenges to Accomplishing Enterprise Application Integration (EAI)
- Systems developed in different programming languages.
- Data may live in different database formats.
- Systems may run on different operating systems.
Can lead to a complex mesh of connections & interfaces as each system establishes communications with the others in the enterprise.
Enterprise Application Integration (EAI) via Mesh
Framework (or software architecture) that forms a middleware layer, enabling integration of systems and applications across the enterprise.
Enterprise Service Bus (ESB)
Addresses the problem of inter-systems communication & allows sharing of data & business processes across connected apps or data sources.
Enterprise Application Integration (EAI) via Enterprise Service Bus (ESB)
Method used for implementing EAI
Enterprise Service Bus (ESB)
Acts as the mediator for the interaction and communication between software applications.
An Enterprise Service Bus (ESB)
Set of software development principles for designing and developing software as services.
Service-Oriented Architecture (SOA)
Well-defined business functions that are built as individual pieces of code with interface “hooks” or APIs.
Services within a Service-Oriented Architecture (SOA) environment
4 “V”s of Big Data
- Variety
- Velocity
- Volume
- Value
3 Characteristics of Big Data
- Data volume
- Processing complexity
- Multiple combinations of structured & unstructured sources
Meaning of ‘Variety’ in a Big Data context
Data is not in a single format; email, tweets, images, standard relational DBs, text files, audio/video files
Meaning of ‘Velocity’ in a Big Data context
Data is generated or changed very quickly
Meaning of ‘Volume’ in a Big Data context
Huge quantities of data
Meaning of ‘Value’ in a Big Data context
Big Data is used to solve specific problems or to answer a specific question
True or False: Big Data is solely characterized by large data sets / sizes.
False
Big Data also focuses on processing complexity as well as combinations of structured & unstructured data sources.
Four challenges of Big Data applications
- Can lead to app sprawl or VM density issues that can bottleneck VDC compute resources.
- BC/DR, retention, & security become unmanageable.
- Increased data load rates can bottleneck the VDCs network.
- Typically computing that requires high-speed networks - 10GbE
Components for Handling Big Data
- Inexpensive, reliable storage.
- Analysis tools for unstructured & structured data.
- Hadoop and MapReduce suited to support distributed computing on large data sets.
Typical percentage of historical data in most production systems
60%
some organizations see upwards of 95%
True or False: Applications that load large data sets and run analytics typically require 10Gig Ethernet.
True
Why can using business intelligence applications on VMs be attractive?
Allows for greater utilization
Typical example of how business intelligence applications are run in a virtual environment.
- Business intelligence app run analytics periodically (e.g., to generate a monthly report).
- VMs can be started, the data analyzed, and a report generated.
- Then the VMs can be shutdown and the resources released.
True or False: What constitutes a “large” data set is relative and can vary from one organization to another.
True
A large data set for one org may be 100s of gigabytes.
For another org, it may be hundreds of terabytes.
Definition of a “large” data set
A large data set can be defined as the point at which the organization has difficulty managing the data set.
Applications that:
- scale at an exponential rate;
- example: social networks;
- can handle rapid growth without bottlenecks;
- commonly NoSQL databases
Web Scale Applications
Type of application that is deployed and managed across a large user base
Horizontal Application
Examples of horizontal applications
Desktop applications
Applications that will impact many users simultaneously
Horizontal applications
Example of a horizontal application impacting many users simultaneously
An anti-virus update being pushed to all user desktops and images within the organization
Enable BYOD and, to some extent, desktop virtualization
Mobile apps
Application type that performs a specific task
Mobile Applications
Two basic types of mobile apps
- Native apps installed on the end point.
2. Web apps
In a Mobile application context, web apps are sometimes disguised as what?
Native apps
Applications often architected as traditional 3-tier applications
Mobile Applications
Generally written for a specific device or a target end point operating system
Native applications
Reside on a server and are accessed via the Internet or cell phone network
Web apps
Characteristics of web app software
- Written using HTML and supporting technologies (e.g., XML & CSS).
- Interactive parts are written in Java, Objective-C, or other language as the end point platform dictates.
Traditional 3-tier application architecture
- Presentation layer (the end point)
- Business logic layer (middleware)
- Data access layer (data store or DB of some kind)
6 Benefits of Desktop Virtualization
- Ease of management & security
- VDI image library
- Heterogeneous OSs
- Snapshots
- Green
- Persistent and non-persistent desktops
Two types of virtual desktop solutions
- Server Hosted
2. Client Hosted
Examples of Server Hosted virtual desktop solutions
- VDI - stateful and stateless desktops
2. Remote Desktop Services (RDS)
Example of Client Hosted virtual desktop solution
Desktop running as a VM on client system: full virtualization and paravirtualization
Not the typical VDC deployment
Client Hosted Desktop Virtualization Solutions
Abstracts the OS, applications, and associated data from the PC.
Desktop virtualization
Most common desktop virtualization solution
Server hosted implementations (particularly VDI)
Desktop virtualization solution where the end users instantiate a session on a remote host machine
Server hosted implementation
Require a virtualization platform be installed on the local system.
Client hosted virtual desktops
Offers the ability to switch between several desktops - such as between a Linux environment, a Windows XP corporate environment, and a Windows 7 environment.
Client hosted virtual desktop
True or False: Both server hosted and client hosted desktop virtualization solutions require software installed on the client.
True
True or False: Remote and client virtualization software is available for Windows, Mac, and Linux
True
Benefit of Snapshots in a desktop virtualization solution
Ability to recover or rollback a user’s desktop.
How do desktop virtualization solutions contribute to a greener IT environment?
Thin clients have reduced power needs.
Consolidated computing in data centers can use power more efficiently.
Reusing older end points reduces technology discarded in landfills.
Allows flexibility for stateless images to be used temporarily
Desktop state
Two examples of how desktop state can be used
- A client for visitors that can be rolled back each day.
2. A stateful image where employee changes to their desktop image is preserved.
5 Key Considerations for Application Migration to VDC
- Flexibility
- Elasticity
- Efficiency
- Multi-tenancy
- Costs
Three things that enable flexibility in application migration to VDC
- Templates
- Cloning
- Snapshots
Enables quick scale out of application clusters to address burst and growth
Cloning
Reduces implementation and configuration times
Templates
Four ffficiency benefits for application migration to VDC
- Better utilization of CPU, memory & bandwidth
- VDI
- Lower upgrade risks - snapshot rollbacks
- Lower application testing time
Two benefits of Multi-Tenancy
- Failure isolation
2. Improved resource utilization
Applications which cannot be migrated to VDC
- Appliance-based applications
- Applications with specific hardware requirements
- Application that are not supported by the vendor when running in a virtual environment
- Applications with data requiring isolation
Three examples of appliance applications
- Storage controllers
- Network devices
- Highly specialized software that requires a vendor-supplied hardware platform
Examples of application hardware dependencies
CPU architecture
BIOS
Peripherals and drivers
Three drivers of data / application isolation
- Legal or regulatory requirements
- Political
- Test / dev / production
Why multiple applications’ data requirements should be distilled into a few categories or ranges
So standard data pools can be created
Goal when virtualizing the data center
Reduce (to zero if possible) the number of “special” cases.
Benefit of minimizing the number data classifications
Simplifies management and planning - enabling the benefits of standardization to be realized.
5 data classification dimensions
- Size
- Response time
- Availability
- RTO
- RPO
Examples of RPOs
Zero data loss
< 4 hours loss
< 24 hours loss
Examples of RTOs
Immediate
< 4 hours
< 24 hours
Example availability requirement for email
99.999%
Example availability requirement for Accounts Payable / Accounts Receivable
99.9%
Example availability requirement for Expense Management
95%
Example compute requirement for email
Multiple 4 x 3 MHz
Example compute requirement for Accounts Payable / Accounts Receivable
2 x 3 MHz
Example compute requirement for Expense Management
1 x 2 MHz
Example storage requirement for Email
> 10 TB
Example storage requirement for Account Payable / Accounts Receivable
1GB to 1TB
Example storage requirement for Expense Management
< 1 TB
Example storage response time requirement for Email
< 1 ms
Example storage response time requirement for Accounts Payable / Accounts Receivable
< 10 ms
Example storage response time requirement for Expense Management
< 100 ms
Example network requirement for Email
Global
Example network requirement for Accounts Payable / Accounts Receivable
Intranet
Example network requirement for Expense Management
Local
Example access requirement for Email
Internal and External
Example access requirement for Accounts Payable / Accounts Receivable
Internal
Example access requirement for Expense Management
Limited internal
Example management requirement for Email
Fully automated
Example management requirement for Accounts Payable / Accounts Receivable
Semi-automated
Example management requirement for Expense Management
Manual
Example RTO/RPO requirements for Email
RTO: Immediate
RPO: Zero
Example RTO/RPO requirements for Accounts Payable / Accounts Receivable
RTO: < 4 hours
RPO: 4 hours
Example RTO/RPO requirements for Expense Management
RTO: < 24 hours
RPO: 24 hours
Example Regulatory / Compliance requirements for Email
High
Example Regulatory / Compliance requirements for Accounts Payable / Accounts Receivable
Medium
Example Regulatory / Compliance requirements for Expense Management
Low
Cost if Lost / Exposed for Email
Very high
Cost if Lost / Exposed for Accounts Payable / Accounts Receivable
High
Cost if Lost / Exposed for Expense Management
Medium
An email application can be compared very closely with what?
A utility.
When email fails, business can be stopped.
Six benefits of virtualizing applications
- Multiple OSs can run on a single server
- Reduced CAPEX
- High Availability
- Enhanced Business Continuity (BC)
- Improved Disaster Recovery (DR)
- Centralized Management
Two things that allow applications to be moved from one hypervisor server to another as needed to balance load and adjust for changes (failures) in the underlying server hardware or faults in the application servers themselves.
Hypervisor clustering
Application fault tolerance
How CAPEX is reduced when applications are virtualized
Increased energy efficiency
Less compute
Less network infrastructure
Three things determined by application SLAs
- Amount of processing & memory needed in the compute space to meet the SLA
- How much bandwidth is required
- How much capacity is needed and how the life-cycle of the data is handled.
Network resources needed to support SLA
Bandwidth
Storage resources and technologies needed to support SLA
Tiering
Compute resources needed to support SLA
CPU
Memory
Connectivity
What dictates the technologies employed within the VDC infrastructure?
The applications that will use the resources
5 Challenges with Virtualized Applications
Resource allocation Application visibility Storage management Management complexity VM sprawl and abandoned machines
Resource allocation issue with virtualized applications
Shared environments can starve some applications
Application visibility issue with virtualized applications
Hypervisors monitor VMs, not the applications running on the VMs.
Storage management issue with virtualized applications
Moving from DAS to SAN / NAS
Management complexity issue with virtualized applications
Not only manage the application, but the VM and hypervisor as well
VM sprawl and abandoned machine issue with virtualized applications
Creating new machines becomes too easy
Usual source of first alert of poor application performance
End user
True or False: Most hypervisors have some concept of resource pooling that can detect, alert, cap, or proactively move applications that demand excessive resources to avoid starving other tenants.
True
Much of the mobility functionality of modern hypervisors (i.e., the ability to move a VM and its applications among multiple hypervisor hosts) is dependent on what.
Shared storage resources
Increasing node specifications is an example of what kind of scaling?
Vertical Scaling
Increasing node count is an example of what kind of scaling?
Horizontal Scaling
Increasing an application’s capability to handle more data, more clients, or a broader geographic region
Scaling an application
True or False: Another way of upsizing an application is to decrease the application’s response time.
True
Three ways to improve the architecture
- Faster networks
- Appropriate storage protocol
- Faster physical disks
Migrating an applications hosted on a NL SAS-based LUN to an SSD-based LUN should provide a major increase in performance.
True
Provides a major opportunity to improve response time
Hard disk architecture
Used to solve a specific application problem, or in conjunction with virtual networks, or to address a scaling problem
Virtual Appliances
Encapsulates all application attributes inside a VM
Virtual appliance
Can be developed in-house, obtained via vendor, or community sites.
Virtual appliances
6 Examples of Virtual Appliances
- Load balancers
- Routers / firewalls
- Databases
- Web servers
- Platform stacks
- Desktops
A pre-build VM image
A virtual appliance
True or False: A virtual appliance is a pre-built VM image.
True
Used to quickly deploy standardized applications to solve a particular need.
Virtual appliance
Can be used to quickly and easily deploy specific applications encapsulated in their own self-contained Virtual Machine wrapper.
Virtual Appliances
Benefit of replacing application server clustering with a hypervisor fault tolerance feature.
Standardization
Hypervisor features that can be leveraged to replace application functionality
Fault tolerance and high availability
Snapshot backups
Clone VMs
True or False: There are many features available via hypervisors to support fault tolerance, high availability, disaster recovery, VM deployment, etc.
True
VDI
Virtual Desktop Infrastructure
A set of techniques and technologies that allows remote access to a desktop environment, usually residing in a data center.
Virtualized Desktop Infrastructure (VDI)
Benefits of VDI
Anywhere, anytime, multiple device access.
Standardized desktop.
Ease of managing multiple desktops.
Compliance.
Ease of support.
Centralized BC/DR and backup.
Reduced exposure to environmental hazards and theft.
Promises flexible access from any device capable of running the client-side software needed to connect to the virtual desktop.
VDI
Allows a desktop to reside in the data center and is under IT’s control.
VDI
Cost savings of VDI
Offers relatively long-term cost savings.
Do not expect immediate cost savings.
Important in a VDI design
High Availability
Why VDI outages are serious
In VDI, user desktops become a data center application.
An outage can impact many end users.
What can cause huge I/O loads in VDI deployments?
Normal or trivial tasks, such as:
- Boot storms
- Virus scans
- Indexing
These and other rather mundane items take on a different personality when running across hundreds or thousands of user desktops at the same time.
Occurs when a VDC environment must start up hundreds or thousands of virtual desktop environments at the same time.
Boot Storm
Ways to minimize the effect of boot storms
Proper use of storage tiering, broad network access, sufficient compute resources, & hypervisor/VDI features.
Benefits of Desktop Virtualization
- Ease of large scale management
- Enables BYOD
- Data protection
- Reduce / eliminate environmental factors
- Standardize platform(s)
- Work anywhere
Impacting and transforming organizations’ businesses
Consumer-class technologies & services such as mobile applications, Web 2.0 social networks, multimedia, and mobile devices.
Create challenges for IT departments and blur the boundaries between business and personal use.
Consumer-class technologies & services such as mobile applications, Web 2.0 social networks, multimedia, and mobile devices
Adoption of consumer-grade end user devices, supporting technologies, and the empowerment of choice given to the end user
Consumerization of IT
Traditionally known as Terminal Services
Remote Desktop Services (RDS)
The traditional Server Based Computing (SBC) technology
Remote Desktop Services (RDS), a.k.a., Terminal Services
VDI Benefits
- Recent technology
- Better fault tolerance
- Can deploy thin density
- True user segmentation
Primary benefit of RDS
User density - the ability to serve more user desktops per host CPU core
Drawback of RDS
Sessions are more dependent upon the stability of the underlying OS, resulting in a potential decrease in stability of the end user platforms.
How does VDI result in better fault tolerance?
VDI uses fully segregated VMs which offers additional fault tolerance, such as VM failover and redundancy.
Why is software compatibility with VDI generally greater than RDS?
RDS can have issues due to the single OS running underneath the sessions.
General comparison of RDS and VDI technologies
While RDS is by far the most mature technology, the decrease in storage and CPU costs is helping to mitigate the costs of VDI, leading to a huge growth in that area.
Outpacing the use of PCs
Use of thin clients
True or False: Many thin clients are being designed for VDI.
True
Manages server-hosted remote desktop to clients
VDI Connection Broker
Works in conjunction with authentication / authorization and hypervisor
VDI Connection Broker
True or False: a VDI implementation can consist of virtual desktops that clients can access with end-user credentials.
True
Preferred VDI implementation
Uses a connection broker to handle assignments of virtual desktops to clients
VDI Connection Broker Functions
- Assignment of a desktop from a pool or a dedicated virtual desktop.
- Web interface that can create secure SSL connections to remote desktops.
- Directory Services integration
- Full USB support
- Support for various display protocols
- Integration with Remote Desktop Services (former Terminal Services)
Techniques for separation of different types of data
- Home directories and group policies
- Roaming profiles and folder redirection
- VMware user data disks
- AppSense Environment Manager - Profile Management
- RTO TScale - Application Load Management
Optional, but reduces storage consumption
Data separation
Benefits of data separation
Separating user data reduces storage consumption.
Separating different types of data allows for proper storage policies to be set.
Prime reason for converting to VDI
Increased data security
In VDI, the only connection from outside the network
From the end users remotely controlling their desktop systems.
In VDI, where additional security and authentication is configured
Via the connection broker
Require just as much hardening as physical systems
The virtual desktops themselves
Does not negate the need for true security software to be installed
The temptation to use VM snapshot reversion to solve all virus and corruption issues
Should be part of a segmented end-user network, even though they may be physically hosted in the same rack as production servers, or perhaps on the same hypervisor itself
the VMs hosting the virtual desktops
Keeps data and application traffic on the secure corporate network
VDI
VDI Security on the Physical Client
Use secure protocols to access virtual desktops
VDI Security on the Virtual Desktop
Do not rely on “reversion” for security.
Have same anti-virus, restrictions as physical desktops.
VDI Security - Segmentation
Place user desktops in the same domain as physical desktops.
Segment from protected servers.
VDI Security - Physical Infrastructure
Standard security considerations
VDI is NOT recommended if the organization…
Is averse to complexity.
Does not have a solid infrastructure.
Does not have a budget for it.
Consider VDI if the organization…
Wants to reuse legacy desktops. Supports a distributed workforce. Wants to contain intellectual property. Wants to standardize. Has compliance issues. Has a BYOD environment. Is willing to relearn desktop management.
Can be used with or without VDI
Client-Side Application Virtualization
Access applications when / where / if needed
Client-Side Application Virtualization
Applications are streamed, thinly installed, or presented on the end point machine
Client-Side Application Virtualization
Users can access the latest version of the application without installing the entire application on their machine.
Client-Side Application Virtualization
Administrators have a single maintenance point.
Client-Side Application Virtualization
Applications with different or conflicting requirements can run on one device.
Client-Side Application Virtualization
Various products allow a level of deterministic behavior based on the location of the end point, such as inside the organizational network or a trusted remote location.
Client-Side Application Virtualization
Apps loaded on the end point as they are used
Application Streaming
Ways that applications can be handled in a Client-Side Application Virtualization environment
- Streamed (loaded on end point as they are used).
- Installed onto a lightweight VM on the end point.
- Run remotely accepting keyboard & pointer input while returning video and optional sound.
One of the main benefits of application virtualization
Providing a central distribution point
Advantages of a central distribution point for applications that are virtualized
- Users always have access to the latest version.
2. Provides administrators a single location to apply patches & monitor usage.
True or False: Application virtualization allows different versions of the same application.
True
True or False: Application virtualization allows applications that do not directly support the end point to be run either virtually, or via presentation to the end point.
True
Has tremendous implications to BYOD environments, or environments that wish to prolong the life of otherwise legacy hardware and software.
Application virtualization allowing applications that do not directly support the end point to be run either virtually, or via presentation to the end point.
True or False: Application Virtualization is separate from, and not dependent on, Desktop Virtualization
True
True or False: Application Virtualization and Desktop Virtualization complement each other nicely.
True
Two Basic Application Virtualization Architectures
- Presentation Mode
2. Thin Install
Application Virtualization Architecture where:
- Application runs on a remote machine.
- Keystrokes & pointer sent.
- Video, audio sent back.
Presentation Mode
Application Virtualization Architecture where the compute resources runs on a system that is remote from the end user, like in the data center.
Presentation mode
Application Virtualization Architecture where:
- Application runs on the end point.
- Application packaged to run inside a minimal VM wrapper.
- Application might be streamed.
Thin Install
Application Virtualization Architecture where the software and needed support / config info is downloaded to the end point and executed in a virtual “bubble”.
Thin Install
Isolates the program and the configuration information from the host OS
Virtual “bubble”
In this application virtualization architecture, the end point need only be compatible with the wrapper program that contains the virtual bubble.
Thin Install
True or False: In a Thin Install application virtualization architecture, different versions of the same program that are incompatible with each other could potentially run on the same machine at the same time.
True
True or False: In a Thin Install application virtualization architecture, software that does not support the end point architecture may be able to run on the end point in the virtual bubble.
True
Application and wrapper components are downloaded as needed.
Streaming
Only enough of the program needed to run is downloaded to the end point.
Streaming
Provides an opportunity to employ storage tiering to place this type of data in a pool that had different characteristics than, say, the storage holding the VDI image itself.
Storage of user and data
Requires the use of some kind of software tool that understands the requirements of the virtualized software and can package it for use in the environment.
Thin Install
Terms used to refer to the use of some kind of software tool that understands the requirements of the virtualized software and can package it for use in the environment.
Packaging or Sequencing
True or False: Thin install and streaming applications require a packager.
True
Ideal for horizontal applications
Greater efficiency of resources
Use with/without VDI
Deploying Application Virtualization with VDI
True or False: Application Virtualization is an emerging technology, which can be deployed even without VDI.
True
A natural enhancement to a VDI environment, because it allows more users to connect to simplified and pooled virtual desktops.
Application Virtualization
a standard corporate Windows 7 desktop with only the barest applications required
typical virtual desktop
Location of users roaming profile
NAS
Default documents and desktop storage for the user
Users Roaming Profile
True or False: In an application virtualization environment, since the NAS solution is also in the data center along with the virtual desktop, working with documents is provided at-near or at-local disk speeds.
True
May provide the greatest benefit in terms of managing user images and resiliency against hypervisor failure.
Application Virtualization Environments
True or False: The connection broker works from a pool of desktops, and is not tied to a specific desktop hosted on a single hypervisor.
True
Benefits of Client-Side Application Virtualization
- Cross-platform execution
- OS isolation
- Deployment flexibility
- Ease of license tracking
- Ease of patch deployment
Drawbacks of Client-Side Application Virtualization
- Some applications cannot be virtualized.
- Different and additional management tools.
- Complexity
Application Consideration Checklist Questions 1 - 4
- If this is a legacy app, can app convert from physical to virtual (P2V).
- What is the VDC users tolerance to latency?
- How critical is this application to the business?
- How many users use this application?
Application Consideration Checklist Questions 5 - 8
- How sensitive is the data used by the application?
- Are there regulatory concerns with the data?
- Is the application slated to be decommissioned?
- Is the source code available?
Application Consideration Checklist Questions 9 - 12
- What is the application’s stack? (OS/DB, n-tier, etc.)
- What are the application’s integration points with other apps?
- Estimated storage space need?
- Estimated storage latency requirement?
Application Consideration Checklist Questions 13 - 16
- Is the application usage pattern predictable?
- What are the monitoring needs of the application?
- Have transient applications workloads been taken into consideration (test/dev, migrations, conversions)?
- Can the app be simplified by breaking it into pieces or purchasing COTS or use SaaS to replace homegrown elements?
Application Consideration Checklist Questions 17 - 20
- Can interprocess communications be replaced with more network centric technologies?
- Are apps designed to dynamically scale across all components?
- Are load balancers designed in the solution?
- Can the application interface handle new data stream inputs?
Application Consideration Checklist Questions 21 - 22
- What is the acceptable application response time?
22. What is the cloud tenant / VDC users tolerance to latency?
Applications should be built for what two attributes?
Resiliency
Fault Tolerance
How are resiliency and fault tolerance applied to a distributed application?
Any failed shard or cluster should not halt or bring down the entire application.
Design applications based on what assumption?
The assumption that applications will fail.
High Availability Considerations 1 - 4
- Does the application require High Availability (HA)?
- Is the application designed for HA?
- Is the launching of an application cluster automated?
- Is the application data persistent?
High Availability Considerations 5 - 8
- How are backups managed?
- What are the redundancy options in the event of a hardware failure?
- What are the disaster recovery options?
- What concurrency model is in place for multi-user edits (pessimistic, optimistic, or none)?
High Availability Considerations 9 - 10
- Is the application designed with network-based interfaces?
- Is the application in a shared cluster, shared instance, or dedicated cluster?
Brownfield
Existing systems in place
Greenfield
New installations
Insert or replace into existing environment
Phased installations
Complete replacement of systems
Forklift upgrades
- Maintains parallel systems
- Can be more expensive due to duplicate systems
- Requires some planning
- Whole sections of the environment are replaced at a time
Forklift Upgrade
- May not have parallel systems.
- Can be less expensive, repurpose systems.
- Requires extensive planning.
- Small sections of the environment are replaced at a time.
Phased Upgrade
If the goal of a project is to migrate from a legacy architecture to a VDC, how might a phased plan be broken up?
A phased plan might be broken up across layers - storage, network, compute, and application.
Citrix ICA Client
- Create a VDI infrastructure with app virtualization using XenDesktop and Xen App.
- Infrastructure / App servers & desktops run on separate clusters w/ EMC VNX providing storage for both.
- Cisco UCS used for desktop images
- Cisco Nexus switches for connectivity.
Example of virtual desktop client software
Citrix ICA Client
How can different desktop configurations be deployed for different users based on their role in the organization?
Through integration with Active Directory Services
Provides authentication and authorization, as well as Group Policies for managing the redirection of user data folders and user state persistence between sessions.
Active Directory Services
Function of XenApp 6.5 servers
Streaming of virtualized applications
Function of XenDesktop 5.5 controllers
Control of the desktops
Function of XenServer 6
Hypervisor platform for the VDI environment
True or False: All of the infrastructure components that make up the VDI environment are running as VMs.
True
How many virtualized Windows 7 desktops (approx) can 20 Ciscu UCS B-Series blade servers handle?
~1,000 Windows 7 Virtualized Desktops
Function of Cisco UCS B-Series blade servers
Used to run the actual desktop instances (compute and network)
Function of EMC VNX5300
Stores the VM images for the infrastructure and desktop images, as well as serves CIFS shares for the redirected user profiles and home directories.
Benefits of a VDI infrastructure using EMC VNX, XenDesktop, XenApp, XenServer, & Cisco UCS
- Increased security by centralizing business-critical information.
- Increased compliance, & ability to prove compliance as data is not stored on the end points.
- Ease of managing desktops through standardization & centralization.
- Empowerment of users to work from any location.
- Embraces the consumerization of IT through multiple end point support.
- Increased business agility & adaptability by allowing rapid, flexible desktops to support changes in the business landscape.
Function of EMC Fast VP Tiered Storage
Move active data sources to Enterprise Flash Drives (EFDs) while less active data is moved to a tier more suited to lower performance requirements.
Benefit of tiered storage pools
Allow workload granularity that does not require the DBA to tune the database to meet SLA requirements.
What drives the architecture?
Application Requirements
What needs to be inventoried?
Applications, data, and security requirements.