We call them the Laws of Physics vs. the "Suggestions of Physics" for good reason. 

>>Laws are proven, tested, validated, peer reviewed, challenged, and then accepted as fact after innumerable and honest evidence proves them to be so. We can bank on them.

>>Suggestions are fingers crossed, white knuckled, my promotion hangs on this, Murphy's Law riddled, probably rumor and vendor stained, and oft wrong snippets of information the cost of recovery from such usually outweighs the original benefit we sought by employing them.

The Speed of Light is a LAW

186,000 miles per second.

That's all you get. 

No particle can move faster

Everything has to occur within this limit (oh, and that is in a vacuum) 

-please also see maxwells theory of electromagnetism - it counts too

So who cares? [or why should you]

Well, lets think about this for a second.  Today, the market is running around, sleeves rolled up, getting really serious about virtual desktop [ insert VDI here ]. I dig this, both selfishly and as a technologist and as a heathen for anything new and different. And while everyone is evaluating technologies, solutions, vendors, configurations, use cases, costs, budgets, productivity gains, clouds and other whimsical things - these Laws of Physics effect everything, always, without failure.

Because i have to go mow the lawn in an hour, i will not ramble on into the "everything, and always" comments below...but i will pick on a few.

Use Cases: - The Protocol Incident

This is the first, and most talked about victim of the Speed of Light (or, the transmission of massless particles). Your distance away from the data source, and the latency effects of this distance, and other "hops" along the way can and should bring you in, or push you out of scope for VDI. Then and only then, should you begin to figure out which protocol is best for your use case. This has much more to do what you do inside that workspace, what apps you use, what kind kind of graphics you generate, what type of density you are aiming for (yes, some protocols use more CPU than others in rendering), what kind of pipe you have available, etc etc. Silicon is faster than software and the laws of physics tell us that a GPU 4 inches away from your screen will be faster and better than one in new york, if you are in texas. Its a law, deal with it ;)

Ahhh....the client hypervisor, type one or two - dealers choice. Things local. BYOPC. Helper VM's.VDI Management with local performance - 100,000,000 laptops just woke up because they are now fully in scope. And should be. In fact in the above assessment of YOUR users, create a new column next to CALL CENTER WORKERS called VDI-mobile. The commercial use of client hypervisors is here, and should be.

The good news is all of this is observable.  What do your users have and do today?  There is your answer for what you need next.

Configurations: - Moores Law and when good particles do bad things

This is one i get asked all the time. "How many virtual desktops per core? How many Virtual Disks per LUN? How much memory sharing? Which Hypervisors? Whats on your iPod [ok, some of you are still reading]"

Answer: It depends.

It is a fairy-tale to think that your users will do something magically different when you put them into VDI. They will do, exactly what they do today, and maybe even more of it. 

So, watch them. What do they do? What do they consume? What apps, cpu, memory, i/o, graphics, network, and other unpredictable and volatile things are they up to? Now, picture all of that on more expensive things, aggregated in your data center. 

There are limits. 

There are limits to how many things a cpu can handle, and what average memory access time (AMAT) plays on the performance of a system and applications, and how many random reads and writes you can make to a disk system at an instant in time, and there are mechanical seek time penalties, etc etc etc. Lots of stuff huh? 

Well, the good news is we can observe all of this, plan for it, make smart vendor choices. CONTINUE to watch it, and tune, tweak as needed, being mindful that users will always be Volatile, Unpredictable, and Abusive. This is another Law.

Vendor Choice: - Who buys you a better lunch-n-learn?

Ok, this is a bit dangerous, but to be honest - some vendors are better than others  (gasp)

Some will look you in the eye, and tell you how their products and solutions can help you, and others look you in the eye and tell you why their competitors stuff can't.  There is not much more you need to know on this fact.  Just make sure you have the data, measurements, and limits of physics and profile of user activity in your real world and challenge your vendor to explain to you how their gear can better handle your realities of physics.

Budget: - If i only had a dollar for ...

See above. You typically get what you pay for. And typically, we all buy too much. The power of virtualization is that we are supposed to deliver resources just in time, just enough, ebb and flow resource needs to availability, a magical self healing low admin touch utopia kind of like the Matrix mixed with crayola crayons. I have YET to see this, but hold out hope we get there some day. 

Budget will be determined by use cases, configuration, density, and your individual business needs, and user behavior. 

Please refrain from the tap-out "I can buy a $499 pc argument". We all know you can. And in looking at CAPEX (capital costs) of VDI - i find that kind of analysis akin to running the numbers of building your own personal powerplant to the cost on an electricity bill. If i had a dollar for every time i have been in the CAPEX argument, i would have like $7309 ( 4 times per day, every day, for 5 years - an estimate) Heck, i am not even sure if VDI is about OPEX anymore.

Really.

Do you think of desktops as NOUNS you have to BUY and Maintain, or VERBS you pay to DO ?  Times change.

User Experience: The new Math and Physics of VDI

My opinion, and i am not shy about this, is that VDI is about gains in productivity. If you get them, DO VDI, if you do not DO NOT DO VDI. When we watch the impact to our companies of a $100,000 user becoming 5% more productive because we invoke a new technology, we can calculate the gain, and this gain, is a much a law of economics as the laws of physics that enabled it.

So, User Experience = Productivity.

Productivity is the metric, the new math, to use to determine if VDI is appropriate [ insert hvd, client hypervisors, app-virt, and cloud here ]. CAPEX is dead. OPEX helps in powerpoint to executives, but if you really want to WOW them, throw up a productivity or USER YIELD slide. (explain to them why we moved away from typewriters to pc's in the first place)

The cool thing is we can observe this too. We may also now have the power to proactively and preemptively tweak our configurations to insure optimal user experience, maximum productivity. 

What your users have today, what your users do today, will determine what (if anything) you need tomorrow to increase productivity. There are vendors very willing to help you spend your budgets. 

BUT, before you do this, know how the laws of physics effect the realities of your environment. Your users behavior will determine use cases, those will determine configurations, those will determine density and vendor choices, which will touch budget, and getting what you pay for you may, just may, be able to boost productivity. 

Time to go mow the lawn.

All great experiments are worthless without observation. 

T.Rex

June 2010

 “I have NOT seen the light, clouds got in the way”

So what the heck does a blog title like that mean?  Well to be honest, for the better part of two years –

harkening back to Paul Moritz’s keynote at VMworld 2008 in Vegas – I have been delivering the line

“Cloud computing to me is a lot like unicorns – I believe in them, but have never seen one, so, I

reserve the right to be a bit skeptical.”

Well. I have seen the Unicorn, a few in fact, gathering in small herds, poking their heads out of the forest,

quiet, but there none the less.

So I flip back through my notebook to roughly 2003 and take a look as the commercial inception, adoption,

success, and failures of server virtualization…then flip forward a bit to about 2007 and beginning of 2008 to

the first real chatter surrounding VDI – and there…right there is the light, the Unicorn. One size can and

should fit all – spot architecting for 600,000,000 desktops (in my world) will take an uncommon amount of time,

energy, and effort – and in fact it only marginally better than what we used to do – putting every resource

IN THE BOX.  So VDI takes it out of the box, and we are seeing trends towards, not only take it out of the box,

take it all apart…but, then kind of glue it back together…..close.

Call it Cloud, Grid, IaaS, DaaS (basically anything with two capitals and “aa” in middle and you have it. The

model that will emerge as the sustainable, 20 year end state, for the myriad of technologies we are all working

to cobble together today – customer by customer-project by project-budget by budget.  This model will not work.

It will fail.  Now, there may be emergent standards that arise, winners, loosers, high flyers, also-rans, me-too’s

and other technologies that lock onto the cloud bandwagon – however the underlying reason they will do so

will not be to leverage marketing buzz, it will be for survival.  It will be because it is the right thing to do. It will be

because it works. Vendors and customers alike will benefit – although those not adroit enough to adapt will

perish on both fronts.  Technology came to us all as a competitive advantage – and over 20 years became the

single largest burden on corporate budgets. Ouch.

Today I heard a presentation from a business partner we are going to be working with here at Liquidware Labs 

[we are contributing some data stream monitoring, etc)- a cloud play so pure – I had to actually architect sufficient

vocabulary on the fly to explain my question in the context of their offering, and it hit me, square between the eyes.

There is no reason for a company to ever exclude cloud services of infrastructure, applications, storage, recover,

security, monitoring, desktop, directory services, storage, etc, etc from their business strategy.

Technically we have evolved to the point of ESCAPE VELOCITY. Virtualization, abstraction, memory, density and

networks are fully ready to shoulder the burden that these new services will offer. Business will have increased

uptime, agility, security, compliance, radically reduced if non-present capital exposure, premiums and discounts for SLA’s.

From the desktop to the data center why would you ever buy more than you need….at this moment?   If we can monitor,

monetize, and deliver any computer science metric, in real time, all the time, with interoperability sufficient to migrate

the task from your cloud to mine, and back again – and yes, at a profit or at a net savings – the illusion truly is that there

is something better to be doing.  I am unsure if the private, corporate Data Center (cloud) has much use looking out over

the next few years. Big clouds gobble up small ones.

Now, we call them the “laws of physics” and not the “suggestions of physics” for good reason. There are some

examples I can think of where quantum leaps in tech are still needed for Cloud Dominance…but I tell ya what…the internet

and caching has taught us an awful lot about how to get stuff where it needs to be as, or right before, people need them.

XaaS wins. Clouds, win. The unicorn is real.

As always, brutal feedback and discourse welcome.

T.Rex

May 2010

 I remembered this over the weekend....and thought of its implications in our current VDI "state" - perfecting my re-purposing of other peoples creative analysis on markets, and why we do what we do...

In economics, Reilly's law of retail gravitation states that larger cities will have larger spheres of influence than smaller ones, meaning people travel further to reach a larger city.

The law presumes the geography of the area is flat without any rivers, roads or mountains to alter a consumer's decision of where to travel to buy goods. It also assumes consumers are indifferent between the actual cities. (ahh, this get tricky)

The law was developed by William J. Reilly in 1931.

A plain English paraphrase would be that the balance or Break Point (BP) is equal to the Distance (d) between two places, divided by the following: Unity or Total (1) plus the Square Root of, the size of Place One (p1) divided by the size of Place Two (p2).d is distance and p1 and p2 are the sizes of the places between which the distance exists; the answer will give the distance from p2, also called a break-point. What is the break-point?

As an example: after leaving a store a you remember something that you wanted to buy; it just so happens that you are headed towards an alternative store b. The break-point can be thought of as the point after which you would travel towards store b instead of store a because of its notional "gravity". This would happen sooner, for example, if store b is an equivalent store but with greater square footage, suggesting that you are more likely to go to store b for greater available utility.

This notional gravity can be influenced by a number of things...

So, i think of a number of elements i can weave into a formula like this (complexity, cost, marketing spend, programatic adoption schemas). One flaw, the law assumes you ARE a buyer. Now, i always heard this law described as Two pizza parlors in a strip mall will EACH do better if another pizza parlor moves in, than if they were operating singularly (i guess the distance cited above).

So, MSFT, based on size, certainly could be argued to have the gravitational PULL- however, each and every other pizza parlor will benefit.

Thoughts ??

OK sports fans, get ready with those hisses and boos. Here goes....

"The iPad is the Ultimate VDI Endpoint" - T.Rex Rohrer January 28 2010

Why?

1. User Acceptance - the iPad passes the first smile test - a really cool form factor that end users will digg. Rich and robust local experience for the STUFF we do when we are not doing work. iPad/iPhone familiarity [ i heart my iPhone]

2. Ultra Mobile - small, light, just big enough, i probably will drop it less than i have my laptop. 

3. Always Connected -With 3g and wifi, I am in an always connected state. Offline will be dropped from the dictionary and wikipedia soon anyway, so - good to have this now.(or soon, i think April)

4. Remote App Delivery - Corporate can provision, update, patch, tweak, and entitle applications centrally, and deliver them to those of us that need them centrally. This boosts agility, manageability, drives down costs. This form of "VDI" mitigates some of the burden of OS costs. Again, depends what you need, depends on your config, depends on your Apps, depends on your users. Depends on the end point OS. But -  i used Citrix Receiver yesterday on my iPhone and i gotta be honest...creating a powerpoint in office 2007 on my iPhone (while not ideal in size) was surreal and cool.

5. Remote Desktop Delivery - Pure VDI as i like to call it. The desktop is centrally located. My IT folks provision, manage, and entitle them. Any device, any place, any time, i can view my desktop. Fully windows XP, WIn7, Google Chrome (another prediction i have) desktop, all my apps, profile, etc just waiting for me to port in. Yes, this is a why VDI bullet, But to use the iPad (ok, maybe with a keyboard dock) you have my vote. why have the desktop local.

6. I can now use the word CLOUD and not grimmace - in about 2007 bill gates and steve jobs at the D5 All things Digital predicted there would always be three screens in our life (until flexible size projection technology is perfected) Small phones, workspaces, and larger entertainment screen. Yes you COULD do all your work from one, but we PREFER larger for movies, smaller for convenience based tasks. So, this iPad blurs the lines a bit, BUT - i get a rich and robust local experience for what i PREFER to do, and, ability to interact with Cloud Services - like my IT departments VDI session they have waiting for me. 

7. Cost - i believe the ALL IN wifi/3g version will click in at under $700. So my IT guy can make me super happy, i get all the bells and whistles for non-work stuff, and she can deliver full corporate managed applications or full desktops to my device as she sees fit. 

8. Lack of Complexity [user]- one button and i know how to use it. One ICON and i know how to get to my VDI desktop (which is a windows session and we all have been assimilated to that). So, no training required.

9. Lack of Complexity [IT] - they buy and provision these devices. They stand up VDI farms. Aside from placing one icon on my iPad desktop so i can launch into their corporate controlled VDI farm....what can i mess up? Dont remember the last time i crashed my iPhone (or mac for that matter) with driver conflicts. :)

10. The "dude" Factor - required for the pervasive expansion of any new technology. I suspect the first time i fire up my iPad, click to launch into my VDI Win7 Desktop, use outlook, powerpoint, ftp some files around....i will utter ...."dude!"

Ok, for those of you who know me - ya know I  have pretty thick skin and can take the slings and arrows of disagreement. Those of you who dont...would really really really like to get your opinions on this.

is the iPAD the ultimate VDI ENDPOINT

[discuss]

 VDI C.O.P™

Coefficient of Productivity

The New Math of the Desktop

A [draft] Manifesto

 2009

J.Tyler “t-rex” Rohrer

We are compelled to move beyond the old math of the desktop not merely for reasons of justification of a new wave of available and potentially useful technologies, but rather because often times advancement of science leave us without proper vernacular and computation to accurately describe what we have just created.

Furthermore, we learn and adjust to our new nomenclatures not because they are clever marketing rouses but because they are functionally relevant, and make conversations of our subject matter contextually relevant. In addition, inclusion of new methods of explanation does not always compel us to abandon old ones.

While my economics professors told me that Economists existed to make Astrologers look credible,  when one observes organization after organization struggling to justify what they instinctively know can be accretive to their business, well, one must speak up. Here say I:

CAPEX: The Capital Cost Expenditure to acquire technology

OPEX: The Operational Cost Expenditure to deliver the technology in a useful way

C.O.P™ -  Coefficient of Productivity – the net incremental GAIN any new technology delivers to its users (both end users and administrators)

This coefficient shall be an index, inclusive of (but not requiring) any number of metrics including security, agility, recoverability, availability, ease of management, burdens on the organization (a negative) each representing a fractional percentage of 100% as measured by its impact on the adopter, and INTENDED to be customized for the specific adopter of any technology.  Theoretically there is no limit to the COP that may result from this exercise-meaning some technologies, sufficiently understood and applied, will yield COPs in any, and often significant degree. (think of the introduction of the PC, telephone, fax machine)  Conversely, it is possible that any technology will / can result in a negative COP.  This can be both because of its burden on productivity, or, (and the subject of further exploration) the mis-use, improper deployment, or, erroneous application to incorrect users.

To calculate the true effect of, and in consideration of, any new technology, one must take CAPEX/COP = True CAPEX

Ie)  CAPEX = $1000

        C.O.P = 1.097 (a 9.7% NET GAIN in productivity)

        TRUE CAPEX =  $1000 / 1.097 = $911.57

An interesting effect of this line of reasoning is that one can also use COP to look at the EMPLOYEE YIELD that results from the proper introduction of new technology to an end user (OPEX covers the gains to the admin).

Ie) If an organization can make a non-revenue producing employee who earns $100,000  1% more productive because of any new technology the generative effect to organization is mathematically $1000.  Expanded over hundreds or thousands  of users, this effect becomes meaningful.

Furthermore, in revenue producing employees – if an organization can enable a $1,000,000 revenue-generating employee capable of closing merely .5% more sales, this yields an effective return to the organization of  $5000. (and could also have a contributory effect on the fixed cost exploration above).

Clearly, the methodology previously used in the desktop space that “I can buy a $399 PC  why would I buy a $750 virtual desktop?” is flawed, and actually out of context.  I liken this to “ Why would I buy a $2000 personal computer when I have a perfectly sufficient $199 type-writer?”

A major underlying assumption of this manifesto is that the explored technology is/can be incremental to productivity gain. The absence of  tangible, and measureable gains compels us to abandon potential technologies. (yes, a car with 6 wheels will drive, but not more productively than a car with 4 wheels perhaps)

In addition, comparative analytics are also needed to not only look at the true cost of a new technology-in this case using COP, or, its productivity gains to the organization, but also this relative to the legacy or current solutions deployed.

One must conduct this analysis not in the absence of OPEX (operating) but rather in concert with such observation.  OPEX is often dismissed as “soft costs” and for that matter not taken seriously because such (if any) savings are not always visible, assignable, or, quite frankly, because the math to obtain such has previously been difficult. Understandably we have previously been exploring (at least in the desktop space, however, I suspect in most new technologies) OPEX and struggling to claim the instinctive productivity COP now describes, in those numbers.  Therefore, the New Math could look like:

OLD:

CAPEX + OPEX = TCO (and thus an Internal Rate of Return could be calculated)

NEW:

CAPEX/COP + OPEX= TRUE TCO (however, while we capture the net productivity gain of the technology and incorporate it into a cost analysis – we have NOT accounted for the incremental revenue possible, or burden employment costs avoided, by the technology) * This will be explored in a separate paper exploring employee YIELD under the methodology of COP*

While this paper does not intend to, in two pages, solve the current debate on the gainful utility of new desktop technologies, when one observes logic that can be improved, one is compelled to speak up.

Therefore, before the critics have their way with “COP” and the salvos of slings and arrows fly, I invite any and all for collaborative and constructive debate in the interest of perfecting this NEW MATH of the Desktop.

T.Rex

Standing on a Whale, Fishing for Minnows

Why VDI?

I have spent the better part of the past three years as an evangelist for the business and technical merits of VDI (and all of the associated hosted/streamed/hyper-vised/abstracted/delivered those three letters stand for these days.)

Yesterday, reading a document by Microsoft of all people on Windows 7 Migration, it dawned on me...clear as a bell ringing....why exactly we are all so fascinated with new desktop strategies. Here is my manifesto. Honest and brutal feedback welcome.

What is the Primary Driver for VDI?

• Capex Savings....nope, Capex is for Wimps. This driver will convince you to pay more I am betting.
• Opex Savings...nope, soft costs are cool, but there is more
• Agiligy....important, but nope
• Central Management ... a bonus, but nope
• Security...really important, but nope
• BYOPC .... cool, but nope
• Business Continuity ... valuable, but nope

Why VDI?

So, roll the clock back to 1977. Everyone is sitting at their nice and stable $299 IBM Selectra Typewriter, pecking away, writing memo's, putting them in bins, folding them into envelopes, and moving information around quite well. Doing their jobs. The typewriter not only cost a few hundred dollars...it only needed ONE SKU to support it give or take (a ribbon cartridge), hardly ever broke, and did the job perfectly, lasting for years.

A sales person comes in one day, and says " I have a new thing for you. It costs $2999, it needs about 20 Skus to support it, you are going to have to hire a bunch of people to learn how to work the software and stuff needed to keep it working. It breaks a lot. When it does, you may not know what actually happened, it will last about 3 years, then i will be back to sell you a new one, when i do, it will take a week or so to get them to you. Your users will have to sort of start from scratch. We call it a PC.. How many would you like to buy?"

The salesperson leaves, we say, maybe for some users, but gosh, at that Capex vs. what i pay now, i just can't justify it. Oh, and this is all taking place during the Oil Crisis, during horrible economic times.

And yet, within 5 years, almost every single typewriter on the planet was ripped out, and replaced with this new thing, the PC.

Why?

So, its now 2009. We have $299 PC's, and a Sales person comes in and says..."I have this new thing for you..I won't be able to match the CAPEX you spend now, i believe i can lower OPEX, I believe my technology leads to security, agility, easier management, disaster resiliency, and a host of other good things. Applications may get easier to use, you may be able to buy less of some stuff, but probably more of others. My solution comes in a million flavors, can be complex to configure, definitely will not work for everyone, is new, and ultimately may lead to your cloud dream fulfillment. We call it VDI. How many would you like to buy?"

The salesperson leaves, we say, maybe for some users, but gosh, at that Capex vs. what i pay now, i just can't justify it. Oh, and this is all taking place during the Financial Crisis, during horrible economic times.

And within 5 years, almost every single PC on the planet could be replaced by this new thing, VDI.

Why did we migrate away from the typewriter, to the PC, and why will we migrate to VDI?

Productivity.

If I pay an employee $100,000 a year, and i can restore 5% of the productivity lost by PC "issues" using a new solution = $5,000

If that employee contributes $400,000k in revenue to my organization a 5% increase in productivity = $20,000

If that employee can process 5% more files, logs, cases, calls, complaints....work =  what is that worth?

The current debate has me thinking we are "standing on a whale, fishing for minnows"

VDI is not about Capex. VDI is not about Opex. Those are nice benchmarks to pass smile tests. 

VDI is about Potential Productivity.  What would you pay for that?

J.Tyler "t-rex" Rohrer

September 11, 2009

VDI

www.dabcc.com/media.aspx 

OK, every now and then, we all need to TOOT our own horn a bit. So, here is a new Podcast I recorded for Douglass Brown (yes, a vdi.com patron member!!) that just posted today.

About 54 minutes long, however, hoping you all enjoy!!

T.Rex

Now, this could get pretty interesting

vmetc.com/2009/08/22/vmware-will-virtualize-google-chrome-but-will-chrome-run-vmware-for-vdi/

 viewyonder.com/2009/08/16/feeding-the-it-shriekometer-5-vdi-anti-patterns/

Thanks to Steve Kaplan on TWITTER for finding this... thanks to ViewYonder for a very intelligent piece !

www.pcworld.com/article/170038/vmware_loses_its_magic.html 

Wow...where to begin. In the interest of proper disclosure, VMware did buy my last company and i did work there as a desktop subject matter expert for 14 months, however, that also gave me insight into the inner workings of the company that DEFINED the virtualization market.

No, VMware has not lost its mojo. It is going through the growth curve every company does when it creates radical new technology, spawns a multi billion dollar market, sees competition enter, and defends.

Pundits beat these companies up, because it sounds and sells more than RA ! RA ! RA !

I had a typewriter.

I had a calculator.

I had a PC.

IBM, Texas Instruments, and Dell are all still here.

So, now i have VDI. 

10 Years from now, VMware will be here, MOJO in tact.

And we will be beating up some other new technology innovator who has grown from the garage to IPO.

Just my $.02

T.Rex

So, i read this...

news.idg.no/cw/art.cfm 

then thought this...

If MSFT truly is, and i have heard this on first hand account, not thrilled and uncaring, and maybe even hostile towards VDI, ala their pricing schema for windows Os's inside of desktop VMs .. well, we have all chased the fabled linux desktop like we search for unicorns...so, what if 

GOOGLE ANDROID is the new desktop OS??

Hmmmm..why not?

blog.bobpeers.com/2009/02/12/google-android-in-virtualbox/