Extending applications to remote users can improve efficiency and reduce costs by allowing a more distributed workforce while maintaining security and compliance. This allows customers to develop products globally. However, legacy application delivery methods can disrupt distributed and global product design by not providing visibility into remote environments. Additionally, storage costs often represent 40-60% of virtual desktop infrastructure (VDI) budgets and can be a barrier to adoption without proper planning and assessment of storage needs upfront.
19. “The No. 1 challenge associated with any SHVD [VDI] project, is the lack of storage planning resulting in higher-
than-expected storage costs.” … “storage represents 40% to 60% of the SHVD budget and, in many cases, is a
barrier to entry.”
VDI Cost By Component
Source: Attacking the High Cost of Desktop
Virtualization (July 2013 Gartner ID: ID:G00250192)
Editor's Notes
Take @30 mins to look at challenges faced today – challenges that I’ve come across, which hopefully you agree with – and challenges that you’ve come across and look to understand and address them .
Understanding on who has tried this
And won (and why)
And failed (and why)
And not yet tried (and why)
Undoubtedly here with vendor, indeed work for a vendor and I’ve experience with working with these products in a stack that has pedigree and reference cases. But the purpose of this workop is to work understand business needs, drivers and goals, give insight into how to focus a project on success given common pitfalls
End on Is it all worth it?
More than simply being able to work on a beach… while knowledge workers are high value; range of roles in an organisation that can benefit from desktop virtualization, a range of job functions and capabilities that are delivered.
Also desktop virtualization is more than just “delivering a virtual desktop” – it encompasses yes virtualizing an os to deliver desktop applications; but it is about providing access, remotelky, provisioning services, application delivery. New approaches to user setup and delivery.
IP protection - e.g. off-shore developers / engineers - http://www.thousandtyone.com/blog/ProtectingIntellectualPropertyVsTrustingEmployees.aspx
Data centralisation
Ready access to information,
propagation
o Global workforce
o Security of Intellectual Property
Engineering drawings
Bills of Materials
Cost info
Supplier info
Customer info
Lifecycle data
Product design decisions
Rapid response to business change – e.g. substitution, business acquisitions; Rapidly set up and staff branch offices
Reduced risk – providing disaster recovery solutions; (snow storms, sick child home from school, natural disasters such as the terrible tsunami that hit Japan in March 2011)
Driver: Work from home programs and disaster recovery programs benefit tremendously from centralization and security of design/engineer applications and product design data in the data center.
User location flexibility – working from home, remote locations,
Simplified maintenance of desktop images
Simplification of the end device – thin clients; reduced power consumption
BYOD/BYOC/COPE
Allow wider device choice
Enable access to standard resources and applications
The ability to pull up design documents and sophisticated models on the shop floor or at a customer location is becoming essential.
Key Points:
Now, an engineer can easily confer with manufacturing on the shop floor to get design feedback or share information.
Now the account team can meet with customers at their facility to demonstrate and show designs on laptops and tablet devices and even make modifications on the fly instead of having to back to the office to make changes.
Transition: The net benefit of delivering on these use cases is improved efficiency and operational savings.
Access to legacy applications
o Faster time-to-market
o Cost efficiencies
o Get maximum value for spend#
Drive: Organizations have improved the efficiency and agility of their design & engineering operations, reducing time-to-market while driving down operational costs.
Key Points:
One of the biggest benefits of HDX 3D Pro is that it eliminates the time wasted in transferring large files across the network before a remote worker can start to work on them. Having development teams working productively during different time periods is achievable by centralizing data and making it available to other geographic teams in real time vs bulk transferring over the network files on a daily basis.
A large global manufacturing firm was able to justify the upfront transition costs associated with moving to a virtualized application and workstation environment solely on the network savings achieved from no longer have to bulk transfer 100’s of gigabytes of data to multiple geographic locations daily.
For large automotive companies, as an example, every day shaved off from time-to-market can be worth a million dollars/euros or more, just for a single car model!
Follow-the-sun development ( 24 X 7) development cycles is becoming increasingly important to remaining competitive. Having the ability to share hardware resources across multiple users occupying different time zones is cost effective.
Failure is an option – and many have failed in past./ how use that failure to help your success.
How to fail on a Desktop virtualization project (http://blog.whatwoulddando.com/2012/02/22/how-to-fail-at-vdi/)
Start desktop virtualization not knowing the business problem you’re solving
Assume you’re going to save capital by rolling out desktop virtualization
Rush to rollout the solution
Start a pilot with no success criteria
Have no way of measuring or defining the end user experience in your environment
Assume you know what your users want
Try to virtualize the entire user environment
Have your success criteria be whether you can access a Windows “desktop” remotely
Put the project in the hands of your VMware vSphere administrator
Don’t evaluate Hosted Shared Virtual Desktops (RDS,XenApp,vWorkspace)
Ignore anti-virus
Ignore the user profile/personalization
Use the same images for physical desktops for virtual desktops
For some environments
Ignore the impact to storage, network, datacenter resources
Believe what your hardware vendor is telling you about cpu overcommit
Do user acceptance testing with 5 IT users
Use vendor estimates for IOPs
Lack of focus on goal (or goals) of the project (best projects (on time, happy customer, most useful – with goals; most character building where goals poorly (if at all defined))
What can desktop virtualisation deliver to your business – because organisations, departments, business units can and do have different goals.
Not asking this question is not only putting the cart before the horse, its putting the cart first and then getting rid of the horse.
Key focus on your organisations goals – recommend to group tech requirements into key areas such as Mobility; Security; Personalisation; Devices. Group business requirements into costs, timescales, risks
Key questions should consider
Budget timescales & duration (mergers and acquisitions) skill set, & risk
Aspirations
Constraints
Is the goal to be cheaper? To cost “less” than “traditional desktops?” Not necessarily not out of the box – why?
Desktop virtualisation solutions are replacing something that was often not properly costed in the first place. Tthe end solution often appears to be more expensive than simply replacing existing PCs.
High capex cost – servers+storage+licenses. Yes greater longevity of end-devices - but now have to manage and maintain server infrastructure.
If it is to allow BYOD – what devices are supported, VDI is one piece here; external access, networking, data-at-rest security.
Improve management? Deliver apps faster? No – still possibly to put in VDI with poor management: indeed more components to manage.
Desktop virtualisation platform can lend itself well to better management; better security – and also team with …
communication
Communication (http://www.fiona-campbell.co.uk/blog-details.aspx?hid=19&nid=35)
IT IS GOOD TO TALK
* Start a pilot with no success criteria
Try to virtualize the entire user environment
Rush to rollout the solution
Have no way of measuring or defining the end user experience in your environment
Do user acceptance testing with 5 IT users
1. Stay out of your own head
2. Really listen to others
3. Help others get clarity of thought
4. Keep it simple
5. Make it easy for others to talk to you
In-fighting and controlDesktop virtualization crosses the boundaries of the server group, the desktop group, and the networking group. Who will control it? Who will get mad?
Application migration, discovery and management –
Long tail application deployment;
Dinosaur? Support for legacy applications is often a driver for a VDI project; but in itself can be a significant pain point. Am on ctx user group committee, have attended VMUG and virtual machine user group events –constant concern/niggle/rage that understanding your applications can be a major cause of project timeshift
easy early on (typical desktop suite; pdf reader, common applications by default included in DaaS environments. Harder as the project moves on – Indeed driver for internal VDI – because application understanding and control is in-house. Can get to the point where maybe, just maybe all those applications cannot be virtualized and the goal of fully virtualised needs to be tempered.
More often underestimated the number of applications an organisation has. Communication is key. Initial project might have 50-60 apps. Not unusual to
Discovery of those applications; model plan and deploy trhose applications into a virtual desktop ewnvironment
Tools to do this… appdna, vworkspace, quest?
Also organisations not just using windows applications
Web Apps; mobile apps; Daas - Where is the data located? Local desktop virtualisation wins out as often nearer the resource.
In their book “VDI reality” - Moving Beyond Windows Applications - Native Apps from App Stores
: App store apps are automatically licensed and secured
Vs
Apps can’t access certain things on the hardware or do some things
A typical goal for a desktop virtualizating project is to enable/deliver a more dynamic/streamlined mechanism for deploying/accessing applications. While apple trying to copyright app store its not a copyright thing yet. The main advantage of an app store is that it’s designed to be something that end users themselves see. (Though once they click on an app to request it, it’s up to you to ensure that the app is able to be installed on their device.)
Most corporate app stores can also be integrated into some kind of workflow, so if a user requests an app that’s not available to them, an approval request could automatically be sent to their manager, licenses could be purchased, etc. RES Software (for example) recently released an impressive app store that helps drive IT automation and integrates with range it services, incuding virtual desktop provisions.
We’re also starting to see app stores that cross platforms. HTML5 apps (whether internally hosted or SaaS) can be delivered to any device, as can Windows desktop applications that are coming from RDSH or VDI in the datacenter. Most corporate app stores can also deliver packaged Windows virtual apps (App-V,
ThinApp, etc.) that will run on the client, and several newer ones can deliver native device apps to tablets and smart phones. Citrix CloudGateway Enterprise, VMware Horizon Applications Centrix could certainly be called cross-platform app stores.
There are many bad habits admins have when it comes to traditional desktop management that don't translate well to a VDI project. In many cases, you may not notice those habits, because the ridiculously powerful PCs under the desks have more than enough horsepower to shoulder the load. When you move computing into the data centre and share resources, however, bad habits multiplied by more bad habits become apparent pretty quickly.
For instance, you don't want defrag to run on virtualized desktops. Imagine the workload placed on the shared disk by 30 or 40 machines running defrag at the same time! (Antivirus scans would have the same effect, for that matter.) Clean the image out and make sure you lock down or remove such features from the image so users can't start them.
So, evaluate how you manage your desktops now and look for holes.
Managing the build; managing the application delivery; managing user data.
Use tools used to help communication/transition to manage and maintain the environment going forward
advantage
Poorly specified hardware, specifically storage.
Storage – because users typically need persistent desktops
With discussion on why this is –
Virtualized desktop workloads are unpredictable, volatile, can overwhelm storage and cause severe performance degradation
Need consistent levels of performance
Extremely Write intensive, Mostly random and in high bursts
Small I/O with averages of 4KB block size
Interactive Writes cause read starvation challenges
Actual user data constitutes a very small portion of the overall Windows IOPS
- But it can be done..
Desktops aren’t dead. Are be supplemented – but not retired. Not ceased to be
\Full transition will take some time – in book new VDI reality gabe and brian stated they believed this his entire transition might take a decade: agree. I’s already happening with a pace that can only accelerate.
Still Atlantis computing works with companies who use multiple forms of desktop and applications to achieve this transformation, including SBC/XenApp, persistent VDI, and stateless VDI, as well as new variations and hybrid models.
The rate at which desktops and applications move into the data center will be a directly impacted by the cost, user experience, and
simplicity of the desktop virtualization model. Given the server performance and the improvements in networking the cost of delivering applications and desktops in the datacenter will drop much faster than the cost of managing renewing and deploying PC hardware - making desktop virtualization more attractive over time.
Today, with Atlantis ILIO and thin clients, you can deploy persistent VDI desktops that are faster and cheaper than
physical PCs. As we move into 2014, that cost will drop even more. Atlantis Computing is focused on making storage more efficient by using intelligent software to deliver more storage capacity and better performance with less hardware to all of the workloads that make up the enterprise desktop. And by rethinking how we look at desktops and developing small but powerful new thin client
It is possible for you to live longer and prosper get to work and move those PCs into the datacenter, starting with persistent 1-to-1 VDI desktops
To refer back to the key success points
Understand your business goals
Communicate with your organisation
Assess & manage your applications & the migration and transition – don’t try and change everything
Assess your environment and monitor afterwards
Get your storage right
Creating your field app delivery stack, take leaders in the field
Atlantis - storage optimisation
AppSense – environment management, licensing, security and performance
Lakeside for system evaluation and monitoring
Citrix providing desktop broker & device enablement secure access technologies and wan optimisation.