To be honest...
7 replies, 1267 views
Posted on May 13, 2016
Edited by Gordon on May 13, 2016
Gordon
Registered user
Joined: February 15, 2013

Hi guys,

concerning 4.0 - to be honest, I am more than a bit frustrated:

VMware bought AppSense two years ago, and of course out of your point of view they should have chosen Unidesk. So afterwards somebody decided Microsoft should be the next potential owner, so everything was developed for Hyper-V: Azure, RDHS implementations came afterwards etc. And of course imho MS would never easily buy a technolgy built on Linux.

Version 2.9.4 for vSphere was published months ago last year - and since then there are no updates to the underlying CentOS? Are we really "Enterprise ready" with something like that?

It looks to me as if with version 4 the bride Unidesk now makes herself as beautiful as possible for being bought by the last possibilty:Citrix - We can now create machines to be imported/pushed into PVS - great for all those existing environments (blablabla, what you can not right now).

But to be honest: We always told our customers in the last 3 years: With Unidesk they now could use Citrix VDI without any PV-servers or Horizon without any Composers - and this is going to make their whole setup a lot easier. 

 

So the question is for me: Where are the real breakthroughs in V4 (and I am asking as a partner not a customer)?

Great to have would be:

- Multi tenancy: How great would it be for all of us, to create e.g. a Firefox layer at one point,check it out and to deliver it to 20+ of our customers (of course with the same underlying WindowOS? Without the need of having domain trusts.

- Pre-Built Layers: Even better: Not to imagine the potential this would have for you guys to deliver OS and software packages for a small amount like the Ninite.com guys do. I can imagine this is not easy, but (over here) this would be a huge argument for anybody choosing VDI: I just download my OS/Software updates from the cloud.

- Migration scenarios: How great would it be for potential customers if they know with the Unidesk VDI Management solution, they can choose (a) their way into VDI and (b) they know, the migration of the users to Windows10 in a couple of years will be as easy as 1,2,3? Can't imagine, what a good sales argument this could be...

And to sum it up: This is a very German point of view.

Gordon

 

 

Posted on May 16, 2016
gstrouth
Registered user
Joined: May 19, 2011

They bought App Volumes not AppSense

Posted on May 16, 2016
Chris Midgley
Unidesk employee
Joined: December 15, 2007

Hi Gordon - thanks for the honest, and direct discussion.  I'm glad to do my best to provide some color / feedback and see where it takes us.

First, we have never sought to be purchased.  I understand many people in the industry assume that is the goal of startups (and for many it is), but it is not ours.  We want to build a significant, growth oriented technology business - and that is what we have always been focused on.  Will we eventually be acquired?  Perhaps, or maybe we will go public - either way our goal is to return high value to our investors.  But if you look at the valuations some of those other companies got, most are not impressive.   We instead feel we are sitting on top of an important technology in a massive market that we can make substantial gains in.  That remains our singular focus, and is why we remain independent with around 1,400 customers and growing fast.  Selling out is easy - building a successful company is not.

So why the change in our messaging and focus then?  Simple - to build a profitable, growth company you need to seek "blue water".  Places where you can sail and there aren't storms everywhere.  With the acquisition of CloudVolumes by VMware, the entire VMware sales team became focused on selling EA's to every account, which basically made simple layering free.  We win against AppVolumes all the time, but we found it much easier to add value to Citrix XenApp where there are large deployments and substantial pain, to enhance Microsoft RDS/RDSH on Hyper-V/Azure, and to expand beyond just View.  That is why we shifted focus.  It was to complement our product lines so we could deliver value in areas that clearly differentiate against our competition. 

Originally we were going to bring Unidesk 2.x into the Unidesk 3.x code base, but along the way we discovered a radically better approach to layering.  This was the birth of Unidesk 4, and we decided to charge towards that and unify all our product lines on that platform (View, XenApp, MS, etc.).  This resulted in Unidesk 2 being the View product, and Unidesk 3 being the RDS/RDSH/Azure product.  And it also meant a lack of innovation on those product lines for a while.  So let's take a look at Unidesk 4 and hopefully you will see why we felt this wait was worth it.

Unidesk 4 is paradigm changing.  Yes, it comes with changes, but it also presents substantial new capabilities that nobody else can do.  One such change is that we no longer provision images.  But today there are far better tools for provisioning than when we started this business... but much more importantly we have been told over and over by enterprises that displacing PVS is a no-starter, don't pass go.  Now that we support PVS, and enhance it by eliminating all the gold image management pain (and then further expand XenApp with Elastic Layering), we are succeeding in landing some of the largest enterprises in the world and believe we will win in the majority of them.  But of course we don't support just PVS, we also support Composer/Linked Clones, Instant Clones / vmFork, MCS, PVS, native storage, Azure, Amazon, Nutanix, and much more.

Some of the breakthroughs you will see throughout 2016 include:

- Support every hypervisor, cloud, broker and provisioning system, supporting/enhancing all existing investments and enabling a simple on-ramp to the cloud

- Support VDI and SBC (XenApp) from the same layer management system

- Eliminate gold image management by merging layers into complete, whole gold images.

- Manage apps and OS images once, and share across hosts, deployment technologies and clouds

- Deliver on-demand apps and persistence (user installed apps, etc) on VDI and on Session (securely to individual users using Session Containers)

- Scales to the world’s largest deployments without introducing any new infrastructure (no clustered SQL, no load balancers, etc).

- Comprehensive DR and backup, including active/active site failover support

- All managed from a single, simple appliance (no more CachePoint applications per host). 

On the topic of the appliances (since you mentioned CentOS), this single appliance can be disabled or turned off, and no operations stop.  Create desktops/session, change app assignments, manage desktops - the appliance is only needed to update the bits in a layer and/or to create a new gold image.  So yes, right now it is CentOS (that is likely to change for different reasons), it is really used only as a simple micro-kernel on an embedded appliance as no systems integration or external management is required.

There is more coming as well.  We have an exciting roadmap through 2017 that is built on this core new foundation.  It includes comprehensive multi-tenancy, and the ability to publish and share layers like an app store (no, we won't be that store, but third parties are welcome to create a business around that model).  It also includes a new management system, with HTML5 UI and REST API based backend ... all of which will be released into open source to enable anybody to integrate and repurpose the technology as they see fit.  And there are some other big things coming... but some of that we need to keep quiet on a bit longer for competitive advantage reasons.

I'm sorry that it has been a long time since we have released major innovations on the vSphere View VDI platform, but as you can see the Unidesk 4 platform has been in development for a long time.  We are now on a rapid release cycle (monthly), and you will see these innovations and many more popping out and unifying our product lines.  Even though 4.0 went GA last week, it isn't ready to displace 2.x or 3.x in most accounts.  We should have full functionality from 2.x and 3.x implemented into 4.x by the end of the year, and that is when we will start migrating all existing customers over.  Even so, we will continue to support 2.x and 3.x deployments for a long time, so customers can migrate on their own schedules (and yes, tools will be available to perform migration including migrating all OS, app and personalization layers).

Again, I appreciate your honesty and directness.  I hope this helps.

Chris (CTO/Founder, Unidesk)

Posted on May 25, 2016
Gordon
Registered user
Joined: February 15, 2013

Hi Chris,

To be honest, we are indeed a bit more direct over here - but what I really appreciate is the fact that the CTO of Unidesk is responding like this to a post in a forum. And this by the way something we are always promoting to our (potential) customers: You are not one of <100m  and the forum is well supported...

I really got your point: And we had some discussions with large potential customers so far (2-25k desktops) as well, they are using Citrix right now, and I completely understand why you guys switched into this direction: They do know that PVS is working and they know that the profile management is not "enterprise ready". And of course it is a huge step for those guys to rely completely on a software from a startup - when beeing CCE-V or something like that.

What is itching us is the fact that we actually lied the last half year in all of our 70+ presentations we have had with potential SMB customers: With Unidesk you can do provisioning, this and that and they will add some great features in version 4.  

It is always great if a company adds some new features to a product. But what I do not understand is the fact that you just "do not provisioning anymore", and there was no need to tell your partners earlier about that fact? I went through all the emails we received the last year - not one word. We heard some rumors about elastic layering and so on, but that was it.

Perhaps we are just not on the right distribution list: But if we are - as a Unidesk partner it would be great to know, what is in your pipeline. Of course I really understand why you would not announce all the new features months before going live.

But it would have been great to know (if possible years ahead) if/why you are going to remove features in upcoming versions.

Thanks,

Gordon

 

 

Posted on May 26, 2016
Chris Midgley
Unidesk employee
Joined: December 15, 2007

I see your point ... that the loss of the integrated provisioning is a "feature reduction".  We are continuing to support 2.x and 3.x for several years - so you do have time, but I know that isn't your point.  2.x continues to be more functional than Unidesk 4, and will continue to be so until around the end of the year when all core functionality will be available in Unidesk 4.  For many customers, we continue to recommend deploying Unidesk 2.

I see now how the unexpected loss of core functionality, that you have been selling value around, is difficult for you and erodes the hard-earned trust of your customers.  I think the big difference here is that we haven't been viewing provisioning as a core feature, but rather an unwanted side-effect of our technology to be able to deliver layering from the OS up (because of how it breaks features in vSphere and View, how it limits scale, etc.)  We felt that there were so many better solutions available today - including direct-to-storage - that the change would be better for most everyone.  But clearly that is not the case for you, and for that, I apologize.

It's difficult to make big leaps forward without changing some fundamental assumptions with new technology paradigms.  Unidesk 4 is a major leap forward, and honestly provisioning isn't the only "reduction" that is coming.  One that comes to mind is that we will no longer provision a dedicated VM per user.  For most we expect this to be great news - because you will be able to share pools of floating, non-persistent desktops across all users, and then elastically attach the persistence/personalization and apps for individual users.  This can reduce hardware requirements, improve overall reliability, and dramatically improve scale.  But it comes at a cost ... some applications can only be delivered in the base image when the machine boots (great examples are anti-virus, identity managers, many device drivers or system services) which means this class of application cannot be elastically delivered.  Unlike any other solution, they can all still be managed in individual layers, but the final "layered image" (VMDK, VHD, etc.) that we create (and the one that is then shared across the floating pool) has to have those more complex apps attached to it rather than delivered elastically.  So for some of the more esoteric use cases, it's possible that there will need to be more pools to meet the various application combinations ... and in some cases, use cases that simply won't work with Unidesk 4 because some user installed apps may just not work since those are always in the persistent layer and therefore are always elastically attached.  Now having said all that, we know how to improve on this and fix nearly all of these use cases - but that doesn't fix in our backlog for releasing in 2016.

Another change is that the concept of managing desktops themselves - assigning apps, removing apps, etc. - is no longer in the Unidesk console.  We decided to change the paradigm to instead build on Active Directory.  You can still do per-user app assignments, but now you use AD group memberships rather than the prior Unidesk wizards.  Is that a functionality reduction?  You will have to decide, but I think it is a major improvement in management at scale ... but some, especially smaller customers, may feel otherwise.

That's all I can think of off the top of my head right now.  But as I've offered before to all our Unidesk partners and customers, I'm glad to schedule a private one-on-one briefing and discussion to go over the entire Unidesk 4 vision / roadmap.   If interested, send me an email and we can schedule it.  My email is easy ... just chris (at) undesk.com.

Chris

 

 

 

Posted on January 9, 2017
Gordon
Registered user
Joined: February 15, 2013

Okay Chris,

I should have made you accept a bet on this one.

We have invested about three years a lot of work in upcoming projects as one of the five or six Unidesk partners in Germany.

Now we are being degraded to be one of the 2.500+ Citrix partners in Germany.

Good job - really well done.

Thanks for not informing us (the partners) in advance again..

Gordon

Posted on January 9, 2017
Chris Midgley
Unidesk employee
Joined: December 15, 2007

I'm sorry I have disappointed you.  Never my goal or desire.  Business is complex, and we have tried hard to do the best we can, to take our customers and partners needs to heart, to build a culture of compassion and customer-first always.  It appears we have failed you, and for that I am sorry and sad.

Chris

 

Posted on January 9, 2017
forbsy
Registered user
Joined: February 19, 2013

"First, we have never sought to be purchased.  I understand many people in the industry assume that is the goal of startups (and for many it is), but it is not ours.  We want to build a significant, growth oriented technology business - and that is what we have always been focused on.  Will we eventually be acquired?  Perhaps, or maybe we will go public - either way our goal is to return high value to our investors.  But if you look at the valuations some of those other companies got, most are not impressive.   We instead feel we are sitting on top of an important technology in a massive market that we can make substantial gains in.  That remains our singular focus, and is why we remain independent with around 1,400 customers and growing fast.  Selling out is easy - building a successful company is not.

Uhhhhh....That last sentence...

Login to post comments