Up Close and Personal with End Users

This past week I went into project mode to assist with an application/desktop delivery plan and design in one of our healthcare customers.  My objective:  interview end users to help understand the pain levels to develop the right solution to deliver IT services for them so they can focus on their own objective:  Taking care of patients.  

Short version of this is, I will sit in front of workers from all over the hospital and allow them to take out their computer frustrations on me...  And I tell you what, I walked in feeling good about it and left feeling even better.  Coming from the depths of the data center, it can be rare that we get out in front of end users and chat it up with them.  We tend to leave that to the analysts or IT liaisons that work with the line of business.  That might have worked on initiatives where we were working replacing the core network or even some of the server virtualization solutions we are deploying.  The end user doesn't really care how we do it, just as long as it works.  And when we do it right, they don't know anything happened at all or it is just a minor blip that won't affect them.

The game changes when we start messing with that little space heater called a PC that sits under the desk. End users notice when that doesn't work, it is slow, it is replaced.  Now start talking about pulling that PC in the data center with virtualization, we are setting ourselves up for a major beat down if we don't address it the right way.  This isn't my first interview session with the end users as it relates to desktop delivery, I have done many of these before.  But it was one that I expanded the scope on and wanted more input and face time with all areas of the organization without their IT teams in the room.  It made me glad I did and I noted the following items:

1. 3rd Party point of view - I believe the users opened up quite a bit to me, it was only me in the room with them for most of the interviews.  The only other person was another member of our Varrow team, and I told them to let us have it in black and white.
2. On or Off Topic, it is all good data - the title of the meetings for them was "Experiences in Technical Environment", and I'm glad it was.  I wanted to hear it all, from applications to printing to badging in and out of the time clock system. Healthcare is very unique in how it's end users are typically very mobile, work all hours (24x7) and when things don't work as planned, lives can really be at stake.  Listening to all their stories really brings that to the front of mind that what they are doing is really a special thing.  It makes us want to fight even harder for them to get the right technology solutions in to make that PC work for them.
3. Communication, the face to face kind - in our busy world of email, texting, tweeting and everything else, we can get overloaded with data. It is nice to sit down in a room with people and get to know how their daily work life is as it relates to using a computer.  I also used a notebook and pen to take notes, not the iPAD or my laptop.  (My hand did hurt, I will admit...) to take the computer element out of the room.  This did include going to remote facilities and putting some focus on the extremely mobile EMS/EMT unit.  They present a use case that can benefit from a cached solution and require unique ways to get data to them in rural areas.
4. Sign of Good will - in the end, getting feedback before a solution is put in place will build some much needed good will between IT and the users we serve.  Consistent surveying can help, but only if you are able to do something with the data collected and they see results.  

Now the thing to keep in mind here is the need to make sure to dig deep on questions.  "My PC is slow" isn't a good answer, be aware of the surroundings and know how to cross examine that answer the right way. You might be surprised how much end users really do know about the systems they use even when they tell you they are "Computer Illiterate".  We are getting more and more tech savvy users all the time, it is the nature of how our kids are being taught in school, the phones we use, how we check our bank accounts. 

As I have stated before, the approach is so important in comparison to the technology.  We definitely need the technology to be on par to provide the features and performance needed, but we won't know any of those needs unless we know how the end user works day to  day.  Healthcare can be very different from other verticals, but no matter what industry you are in, paying a visit to your end users from time to time can be valuable when designing how to deliver the desktop to them.

Getting our Applications to Become Virtualization Friendly

In an earlier post, I brought up how we are taking what we know as VDI and approaching it as a Desktop Management strategy, that encompasses six layers that make up a desktop. While this is not necessarily a new concept, we still don't see it being put into practice when solutions are being proposed in environments we are in.  I want to pick on the application layer today.  This layer gets tons of attention and has for as many years as I've been in this business, but still needs some TLC so we can get our desktops in a more dynamic state and more apt to allow virtualization to be embraced on the desktop.  Application Delivery through a virtualized means is a key component to realizing a positive management experience and a better ROI when leveraging virtualization on the desktop.  The great thing about it is you don't have to be hosting a desktop on a server to use it.  You pull that layer out of the O/S and it can be used on physical/virtual/Windows XP/Windows 7, etc.   

I rarely walk into to a sizable environment where they have not used Citrix or Terminal Services.  Citrix can present applications to an end user from a central location so that it appears they are running local, so we get some great central management functionality with that. The thing that pains me is seeing it only being used on a very small portion of the overall applications in use.  Many times it is in place just to deliver two or so applications to remote users or to one particular use case.  Now, I'm a fan of Citrix XenApp, but many of the organizations I see are in the "we tolerate" it mode.  Many also don't know that the XenApp of today operates differently than the XenApp Server of yesterday.  So the first step we like to take is understanding how to leverage that investment to accelerate the desktop management strategy.  It may be just using what they have more or putting them on a roadmap to upgrade and implement new features.  It many be looking at a different technology such as VMware's ThinApp, which does a great job of simplifying the packaging process and can provide quick results for some common applications like MS Office.   

Now to something I feel is very important to our success.  As many of us have seen with virtualization efforts, application support has been slow to come around.  Desktop or client side applications have normally been put on the back burner behind getting our server apps up and running on a virtual platform.  Numerous business application vendors have given a positive nod to Citrix Presentation Virtualization as a delivery method since you are still running the application on top of the operating system.  When it comes to putting an application in a bubble, we are back to the normal "not supported" statement.  This appears to be changing as we successfully package more applications and as the vendors (i.e. VMware) work closely with software companies on certification programs for virtual appliances (i.e. VMware Ready).  With this added support, next put in a solid approach to assessing our application needs, what our environments are performing like and what kind of users we have.  In my opinion, all of these efforts should lend to better collaboration, integration and applications that are virtualization friendly.  

What Is Good End User Experience?

When we discuss desktop management and delivery using virtualization, the various vendors in the space like to bring up end user experience as a selling point.  Matt Hensley discussed this in his blog and gave some great direction on some of the tools out there that have focused on monitoring and placing a metric around end user experience when VDI is in use:

http://matthensley.wordpress.com/2009/06/08/vdi-monitoring-and-the-end-user-experience/ 

I want to continue to build on that conversation and add in some of the approach details we have seen to be successful as we have started to tackle the desktop.

The first thing we have to make sure is clear: VDI is a subset of an overall desktop management/delivery strategy.  Since the desktop has multiple layers, our approach should not focus just on putting a desktop on a server.  We need to look at how we can break the desktop into manageable layers, then deliver those layers in a method that the use case calls for.  With that in mind, we need to approach the end user experience the same way.  Each use case will call for a different type of experience based on the way the end user needs to work, the applications they use, their environment and location.  Once we determine these items, this will feed into the overall design and deployment path we need to take.

Why do we go through this trouble?  If money is no object and you have limitless network, storage and compute, you could probably get away with just picking the heaviest use case and using it to design the infrastructure for everything.  But we know that isn't possible.  Much of the debate has been around the financial viability of leveraging desktop virtualization.  But many times that is because we just focus on VDI and not the approach of how to delivery a more efficient desktop, and all of it's layers. 

In the past, we haven't had the ability to measure the inner workings of the desktop environment enough to figure out what the users need to get the experience they need. There are many desktop management tools that can give us numbers, but they were not created with virtualization in mind.  Those times are changing, Matt discusses tools that we can use to look into the environment and monitor key metrics to put a score around end user experience so we can design, deploy, optimize and grow the environment.  Some of those tools like Liquidware Stratusphere help us to assess before we even start.  We can group our users, assign the kind of experience they need and design the environment to meet those needs.  We can't promise it will be an easy process, but it normally is a process that can give us the best results.

Enabling the Future State of Healthcare IT

I had the pleasure of attending and presenting at the North Carolina Spring Conference this past week in Greensboro, NC.  My topic of conversation was around "Leveraging Virtualization at the Desktop to Benefit Patient Care" (Link to presentation).  This was not a presentation on VDI, but how to deliver a desktop environment that fits better to healthcare providers needs.  It was a well received topic of conversation and something that is becoming more of a realization with the technology vendors, both application and datacenter based, maturing and building with the end goal of creating an "as-a-service" for our healthcare providers in mind.

The keynotes were the night before and they were on some very interesting topics on some of the trends in Disease Registries and the Health Information Exchange (HIE).  They discussed items such as National Health Information Network (NHIN), Centralized Electronic Health Records (EHR) Architectures and Biomedical Informatics. While all things that were out of my specific area of expertise, all were very enlightening at the developments that are coming about in our healthcare system.  Outside all of the medical jargon, terms and techniques, it was very clear that Health Information Technology (HIT) will have to be up to the task in building the foundation to enable these initiatives that are going on in the industry. Leveraging virtualization and specifically the virtual private cloud are critical to creating the kind of dynamic environments that will be able to house and maintain the amount (and kinds) of data that will be flowing in and out.

Security and continuity of this data are very important.  While security is still a large debate from a cloud perspective, there is a lot of attention being given to ensure this is not a detriment to moving forward. Solutions from EMC such as RSA Archer assist with compliance monitoring of our clouds, both private and public.  

From a continuity standpoint, I believe we can start to nod our heads in the same direction that high availability is being addressed full steam ahead by numerous vendors.  Two weeks ago at EMC World, we saw technologies in storage from EMC such as vPlex and Data Domain deduplication that feed into building a rock solid, efficient and always on infrastructure.  These technologies coupled with the advancements in high availability computing technologies such as Cisco Unified Computing System (UCS) and the features that are native to VMware vSphere provide the tools necessary for the private cloud.    

In healthcare, this is critical not just to supporting the iniatitives discussed earlier, but in benefitting patient care.  Which in the end, is really the end goal in mind.

Desktops from the End User's Seat

Sitting in multiple client meetings this week on the topic of "Desktop Virtualization" it was good to see that many were starting to focus on the key elements we have evangelized about with them over the the past year.  One item that stood out was focusing very hard on overall end-user experience and using an approach to measure it.  Measurement is necessary so we can align the right solution and make that experience better than before.  At the end of the day, building the right solution and being an enabler is what this is all about.  We can have all the bells and whistles we want in the technology, but if we can't position ourselves to sit where the end user is sitting and appreciate what they need, it will all be for not.  

An example of these end users are patients in health care facilities.  They may not be touching the keyboard and mouse, but they are directly affected if a PC is being used in the process of their visit and that PC is not performing as it should.  This may cause them to sit in registration longer and not get treated as fast or even worse, get treated for the wrong thing.  

So how do we make sure this experience is the best we can make it?  The list would be very large, but here are a few things that can help:

• Measurement - assessing what a desktop is doing day to day is critical, and not just performance. Looking at how fast a user can login, open apps, how those apps interact and how fast they communicate.  How fast data comes up and is available for the user to make good use of it.  All of these thing can be captured and then analyzed to put a figure around what may or may not be acceptable behavior.  
• Foundational Stability - A stable technology platform won't necessary make logins faster, but they will make sure they happen with 99.999% reliability.  This doesn't just mean beefy servers and storage, but the right servers and storage built to handle the platforms they are hosting. If virtualization is in play, use servers, network and storage that were made with that in mind.  And never forget about the network outside of the core, at the perimeter.  Even on a LAN, acceleration via either network devices or the right end point device are key.  
• Approach - the process of transforming desktop delivery to make the end user experience better needs to follow an approach that allows for measured success along the way.  It doesn't mean you add in extra complexity or slow the time to deployment, it is there to help keep us all in line.  There are some very cool technologies out there and keeping the eye on the goal at hand can be achieved with a solid approach.

End user applications can't get off the hook here, they need to work better alone and together with other applications.  Building a better end user experience will require applications to be built with the end goal in mind.  This needs to start at development where they take into account the innovative technologies that can be used to deliver the applications and forming the kinds of partnerships they need to make them have synergy.  This is out of most of our control, but there are vendors out there (i.e. VMware and Citrix) that are doing things to try and help the community.  Some application vendors are working hard to offer their products in a web or SaaS format along side of a think client medium.  This is a good starting point that is leading the way to a better application experience.

This isn't a very technical topic I'm covering here (yet), look for each item to be detailed out from a technical standpoint in future blogs.  For more on the foundational stability topics on VDI, check out Matt's blog @ http://matthensley.wordpress.com/.  He covers it in multiple articles on network and storage design.