December 30, 2012 § Leave a Comment
I joined the XenDesktop team in early 2009. In those early days of desktop virtualization, XenDesktop was being bought to solve very specific problems – delivering desktops to outsourced teams in India, on-boarding staff of an acquired company, or for security and compliance purposes.
However, towards the end of 2009, we noticed a trend. Many of our customers were enquiring whether desktop virtualization could help solve the growing challenge of BYOD (Macs, iPhones, Android phones, etc). After the iPad was announced in April 2010, desktop virtualization as a solution for BYOD became an integral part of conversations with prospective XenDesktop customers. Desktop virtualization became the first wave of solutions for BYOD. A customer could virtualize their entire Windows desktop and then deliver the desktop onto any device using remoting technologies.
By mid-2011, however, many customers realized that a Windows desktop experience was not what BYOD users wanted. In addition, the complexity and cost of deploying a desktop virtualization solution, caused customers to look for an alternate solution. Companies started looking at Mobile Device Management (MDM) solutions to lock down the device – phones and tablets. MDM was the second wave of solutions for BYOD. A number of customers have acquired and deployed MDM solutions over the last two years. But increasingly many experts and customers are coming to the conclusion that MDM isn’t the right solution for BYOD either: users don’t want their personal devices locked down and managed by IT. See: What is MDM, MAM, and MIM? (And what’s the difference?)
and The Mobile Enterprise in 2013: Getting Down to Business.
We are now in the third wave of solutions for BYOD. There are a number of different alternatives available:
- Mobile Application Management (MAM), e.g., Symantec, Appsense, Mocana, etc.
- VPN solutions
- Mobile Information Management (MIM) or Document Management Solutions, e.g., egnyte, box, ionGrid, etc.
- Container, e.g., Enterproid, Good, etc.
- Single Sign-On (SSO), e.g., Okta, Ping Identity, Symplified, etc.
Its still early in the third wave, and none of these technologies fully solve the problems of applications and data.
At O1 Works , we want to simplify BYOD by building a solution from the ground-up that:
(i) Works the way your users want to work
(ii) Works the way IT wants to work, and
(iii) Helps create order out of the chaos created by personal devices and SaaS applications
We will share more about what we are building in the coming months.
September 28, 2012 § 1 Comment
Yesterday we performed a small network performance test in our office with the iPhone 5 and iPhone 4S. We turned off WiFi on both phones. Both phones were using the AT&T network. Using Speedtest, the iPhone 4S registered a ping time of 100+ms and a bandwidth of 0.75Mbps. The iPhone5 registered a ping time of 40ms and a bandwidth of 14.5Mbps download and 3 Mbps upload. The LTE network is roughly 20 times faster than the 3G/4G network. But that’s not the only surprising thing.
The iPhone5 performance is very similar to what I get at home on my Comcast cable modem. I saw a report that somebody in New Jersey had seen 50Mbps download speeds. I think the LTE speeds are substantially faster than the DSL speeds in our neighborhood – folks are glad to see a 3Mbps download speed on DSL.
If LTE were more widely available, or even available in my home area, why would I buy cable or DSL internet? Why not buy a wireless router, such as Netgear MBR1515, that connects directly to LTE? Has anyone tried this?
Is this the end of wired networks at home?
May 29, 2012 § 2 Comments
Back in 2000, native apps meant Windows apps. Microsoft built a great set of tools (Visual Basic, Visual Studio) to build Windows apps. Hundreds of thousands of developers were trained on these tools. IT deployed Windows apps to endpoints (desktops and laptops) that they owned. And IT managed the applications with ESD tools (Altiris, SCCM, etc.).
Even in this simple world with one platform, lots of trained developers, and enterprise-owned endpoints, IT still struggled to deploy and manage applications with ESD. I have heard many IT Directors tell me that ESD tools worked 80% of the time. So when they deployed an app, it was correctly installed on 80% of the desktops/laptops. This was a big problem – if you had 10,000 users, then some random 2000 of those users did not get the apps properly installed. IT would then spend days, sometimes weeks, resolving these problems.
I see problems in all three stages with today’s native apps:
(a) Tools: In today’s native world, there are far more endpoint platforms – iOS, Mac, Windows 7, Windows 7, Android (all of its versions), Blackberry, etc. Where are the cross-platform tools that simplify the job of developing across multiple platforms?
(b) Developers: There are very few developers who are trained to build across any one of these platforms, let alone cross-platform. None of the existing cross-platform tools has a critical mass of trained developers? Will there ever be one like all the tools for Windows or Web development?
(c) Management: These native apps have to be deployed onto endpoints that the user owns, not IT. What are the odds that in this chaotic, heterogenous environments the equivalent of ESD – MDM tools – will work any better? Will MAM tools allow similar cross-platform control of these apps?
I can see how the ISVs, SaaS vendors, etc. will hire developers to build cross platform native apps. But will the challenges IT will have in managing these apps on many heterogenous employee owned devices create a backlash against native apps? Or will IT simply give up and stop trying to control these apps?
I don’t see how enterprise IT can successfully build custom native apps like they have been building Windows and Web apps for the last two decades. Too many obstacles of tools and lack of trained developers.
May 28, 2012 § 1 Comment
Frank Gillett at Forrester Research has a compelling prediction: “Why tablets will become our primary computing device?” (http://blogs.forrester.com/frank_gillett/12-04-23-why_tablets_will_become_our_primary_computing_device)
I agree. One of the big reasons I think this will happen is form factor. Frank Gillet has talked about a few other reasons and all of them make sense.
Form factor is a serious challenge as enterprise IT considers how it will mobilize its applications. Most apps will have to be re-written by corporate IT or the ISVs for the smart phone form factor in order to make it truly usable for the enterprise end user. I have talked about why this isn’t easy in a previous post.
Tablets don’t have the form factor problem. An application designed for a 17 inch screen renders reasonably well for a 9 inch one. Most web sites are easily navigated on a tablet. The user experience can be improved to incorporate more touch friendly features, but the applications are usable as they are already.
This makes enterprise IT’s life simpler. They can extend their apps out to tablets, either using a browser or a solution like desktop virtualization, and they can do it all now!
I predict that by 2015, Windows tablets will be a strong #2 in the Enterprise!
May 27, 2012 § Leave a Comment
I left Citrix Systems last month. Its been many years, since a large chunk of my email, contacts, and calendar has not been stored in an Exchange Server. I spent the first two weeks trying to live off a combination of Gmail, Google Calendar, iPhone, iPad, and Mac.
Getting email on all these devices was relatively straightforward. However, I have had a number of challenges with my contacts and calendar. For two weeks in a row, I had to wipe out my contacts and calendars on all my devices, start with a fresh upload into Gmail, and by the end of the day, my 500ish contacts would have grown into 40,000 contacts (with lots of junk) and all the various sync processes would be going haywire.
The app centric model on my Mac was the principal source of problems. Each one of the apps maintained its own copy of data and I had to configure synchronization tasks to move data back and forth between these applications. It was difficult to determine which app “owned” the data and which sync process was responsible for the growing contacts database. I had to shut one off and then repeat the exercise. After multiple iterations, I determined the problem was the merge operation on Gmail. I had merged my Gmail contacts and Exchange contacts using the merge operation on Gmail. It does the merge properly, but puts a lot of junk in the Notes field. This caused all the sync problems because the data wasn’t being processed properly by one of the sync processes. And once one of the databases was corrupted, every other database ended up being corrupted.
In a data-centric world, all the apps work off one set of data – reads and writes go back to the file system. In an app-centric world, each app works off its own set of data, and the relevant data is replicated across all the apps. In a world where we have multiple apps across multiple devices, each app with its own logic for how it consumes data, it is challenging to understand how exactly data is handled.
Apple’s iCloud is a big step in the right direction, but in its current form it exacerbates the problem. It syncs the PhotoStream, but not the photos. It syncs the photos in the PhotoStream, but not the videos. It syncs the songs you buy, but not the songs you upload. In Apple’s app-centric world, they have to do a significantly better job of making sure that everything syncs correctly and as one would expect.
May 26, 2012 § Leave a Comment
One of the most common questions/requests I have heard from CIOs over the last decade – ever since the first mobile data devices showed up in 1999 – was how can you convert an app designed for a 17 inch screen to one designed for a 3 inch screen?
I spent 7 years trying to solve that problem with my first startup – Everypath. We were pioneers in software in a category now called MEAPs. There were more than 75 other companies in the space at that time. We all came up with different tools and models to help people build mobile applications for the new form factors/new devices. Some of those companies still exist – Antenna Software, Syclo (acquired by SAP), etc. The new MEAP platforms provide good tools to build apps for different devices and different form factors. More on MEAPs in a future post.
There were a few companies that claimed that they could automatically convert apps into the smaller form factors. But the user experience of the converted applications were less than desirable. This isn’t surprising. Most good user experience is achieved through a result of careful understanding of the problems one is trying to solve, design-centric thinking, user experience research, and experimentation.
There is no good automatic solution. So to all those folks in IT looking for a silver bullet to help them with their form factor problems: there isn’t one. If you need to target a specific form factor and create an optimal user experience, you will have to use one of the MEAPs to build a new app!
No shortcuts. Unfortunately.
May 25, 2012 § Leave a Comment
Enterprise IT and its users live in two different worlds.
Users live in a world with many devices – iPad, Android phones, iPhones, PCs, and Macs. Some even use Blackberries.
Meanwhile, enterprise IT runs a desktop environment that is 95+% Windows. My guess-estimate is that close to 50% still run Windows XP and IE6, and the only reason they will upgrade to Windows 7 is that Microsoft shuts off all support for XP in 2014.
One big question that CIOs have is how they make the transition from the current enterprise desktop world to the world in which their users live. They cannot abandon Windows desktops, and yet they have to support all these devices out there.
Re-writing apps is out of the question. Many of them are Line of Business custom Windows apps, and too much work needs to be done to re-write these apps to be device independent. Even with all the great cross-platform tools out there, such as Appcelerator, PhoneGap, etc., the amount of work to re-write 100s of applications in a large enterprise is enormous.
They cannot replace all the apps for obvious reasons.
If they had virtualized their apps using tools like XenApp (aka Presentation Server, MetaFrame), then they could deliver it to any device. However, most organizations have virtualized a small chunk of their applications in the previous decade. They focused on apps like SAP that were used by a lot of users and it was far easier to centrally manage and deploy them. So the vast majority of the apps remain un-virtualized.
Newer app virtualization technologies, like App-V, are far too Windows-centric to solve the current problems that CIOs are trying to solve.
It is in this context that desktop virtualization, VDI or TS desktops, solved a critical problem. All the desktop administrator had to do was put all their apps into a Windows desktop (virtualized or not), and then deliver it to their users on whichever device they used. Problem solved. Over time you optimize it by virtualizing more apps.
Three years ago, in 2009, this wasn’t easy or cheap. A typical desktop virtualization deployment needed desktop, server, and network skills. The overall cost of the solution was more expensive than the cost of a standard desktop. It was a whole new management paradigm that had to be learned by a desktop team that had spent the last decade using ESD tools. And yet, it was the best way to deliver desktops to remote users and to all devices.
But today, desktop virtualization costs about the same or less than regular desktops. It is still a new management paradigm and needs collaboration between different groups inside enterprise IT. Luckily, a lot of consultants and system integrators are now trained and ready to assist the next generation of enterprise IT virtualize their desktops.
Paraphrasing Winston Churchill: “Desktop virtualization is the worst form of desktop delivery, except for all those other forms that have been tried from time to time.” Today there is no better alternative than desktop virtualization to delivering enterprise apps/desktops to all users on many devices.