This week I was invited to speak at a “guru forum” of managers and academics who work in information technology. Among the many issues that were discussed, two conflicting trends were identified. On the one hand many corporate organizations are moving towards cloud services and all-in-one outsourced solutions (Oracle, SAP, IBM, …). On the other hand individuals are moving towards a “bring your own” model, bringing their own computers, e-books, cellphones, iPads and other devices to their workplaces. With the advent of smartphones and social media platforms such as Facebook, computing is becoming more consumer-centric and primarily a means for social interaction, rather than just a tool for specific tasks like word processing and accounting.

These opposing trends create a disconnect at the workplace between the ability of firms to manage and control information (especially proprietary information) versus the desire to give employees flexibility and freedom in choosing the tools they really want to use. My view is that the trend towards consumer-centric computing will dominate the other paradigm. There is no turning back the preferences of modern information workers who grew up with their iPads, Android phones and Kindles. Companies should embrace rather than fight the trend.

How do we solve this “me versus you” problem? i.e., organizing information on multiple devices in a way that separates private from work and other shared information in an easy but manageable way? Existing solutions are unsatisfactory because they do not adapt to the different and changing contexts that individuals find themselves in. Companies like Apple, SAP and Oracle take a fully integrated approach, allowing you to run everything on their software and leveraging their own cloud solutions, treating each device as a client. Bringing this to the extreme, you can run entire virtual machines from your own device with everything hosted on the service provider, such as via Amazon S3 or OnLive. Unfortunately this is often an all-or-nothing proposition, so while it creates separate contexts, the operation across contexts is not seamless. You’re basically running separate computers (or syncing to separate clouds) from within your own device, and it is slow and clunky to inter-operate between them.

In contrast, other firms like Dropbox provide services that integrate into your existing applications and folders, but end up being highly fragmented especially when it comes to setting permissions and giving access. Each application and each collaborator needs to be authenticated, so coordination can be a hassle. This week my colleague tried to set up a shared Dropbox folder for the faculty members at our school, and it seemed a lot more of a hassle than it needed to be, especially the bit about inviting each user and trying to get them to actually sign up to the Dropbox cloud.

The good news is that the solution of the “me versus you” problem is closer at hand than many might think. The architecture for such a solution already exists in products like Google Circles and VMware but is not yet pervasive. Here’s an example of what one such solution might look like. At present most operating systems support multiple workspaces, but for now they are all tied to the same set of permissions and applications. Well, imagine a future in which each workspace on your device is authenticated to different sets of applications and clouds. For example, your device could include a personal workspace that authenticates to Apple and Dropbox and which contains your personal files, apps and Facebook page. A second workspace could authenticate to your office, with the IT system at your office determining what apps and cloud services are made available and which of these you can transfer across workspaces. A third workspace could be one created by your friend so that when you visit her house, her workspace would appear along with some of the data and services from her home network that your friend is willing to share with you.

A small number of us already have something close to this setup running on our computers by using multiple virtual machines that are active simultaneously. But it isn’t the same thing. I’m thinking of something with much more integration than is available in existing virtual machines and with much less of the “heavy machinery” that is needed to support multiple operating systems on the same machine (the action is in the data and apps, not in the operating system itself anymore). I also have in mind something more dynamic, for example with the ability to seamlessly add or remove workspaces when the context around a person changes. In the example above, if your friend defines a workspace that is shared with you when you visit her, that workspace should actually exist in a virtual sense, and it should slide on and off your various devices in a consistent manner including your smartphone, iPad and notebook computer.

Granted, the ideal solution in my head might be a bit far-fetched. However I suspect it will become prevalent in the next several years. I don’t know what it would cost in terms of implementation and adoption. However, the fundamental issues are of great concern among industry practitioners such at those attending the IT Guru forum, so I suspect that over the next few years entrepreneurial firms will end up exploring solutions and frameworks along these lines.

[This post was originally published at Core Economics.]

5 Responses to Cloud computing and the “me versus you” problem

  1. tomslee says:

    Thanks for these interesting views. But there are two differences between the world of Enterprise computing and of social computing that you don’t mention, and which I think makes a difference.

    One is the transaction. Oracle and SAP’s solutions are built around transaction processing, and that’s not such a big deal for Facebook or Dropbox.

    The second is the nature of shared and personal information. Much of what you discuss is shared among relatively few people in a pretty unstructured way. But I don’t think that sharing a price list (for example) over Dropbox or even its enterprise equivalent would be very useful.

    Do you think the personal model you describe could extend to these facets of enterprise computing.

  2. Habibullah Khan says:

    I think you have identified very real trends but I believe VMWare for one has a solution already in place. With desktop virtualization which is the final part of their cloud journey (after data center virtualization with chargeback and self service and putting your critical core applications into a hybrid cloud) ; device agnostic access to apps you need with a social element is possible. Most importantly it is policy based so it is efficient and it is secure. VMWare Horizon, VShield safezones and VSphere’s safety features enable this. This is critical because if there is one thing that can destroy the cloud as a trend it is security.

  3. GeorgeK says:

    My concern on clouds is who owns the data if my life changes; divorce, new job, lost job, etc.

    If I can’t pay for the cloud how do I get my data back?

  4. Kaleberg says:

    GeorgeK asks some good questions.

    I can’t quite figure out how the “cloud” solves any of my problems. I don’t want to have to rely on Apple or Google unless they are providing a completely portable solution using open industry standards. Too many of us remember getting trapped with proprietary formats and DRM controls, and we don’t want to have to deal with that again.

    The cloud is basically a big IT department wet dream that lets them go back to the good old days of “big iron”.

  5. Kwanghui Lim says:

    Hi everyone, thanks for sharing your views. I don’t think I have all the answers at this stage, but rather am observing what seems to be a growing trend.

    Tom – Yes technologically there may be a focus on the front end on transaction processing for enterprise solutions (e.g., Oracle). But at the back-end transaction processing is happening anyways. Dropbox, Evernote, etc. are offering a “pretty interface”, while also starting to provide APIs and deeper access that may in the near future end up transaction-oriented too, perhaps even ACID compliant in future.

    Now in terms of shared personal information, I think the usefulness of linking enterprise information may be more useful than we think. It’s just that we haven’t been doing so in the past. Even now I sometimes wished our HR and personnel management systems were better linked to my ability to manage my personal life (i.e., my google-hosted calendars, tripit-hosted travel plans, and other cloud-based information). Putting prices on dropbox might not seem useful on the face of it but given the opportunity, new applications could crop up. For example, for a small/medium sized firm, the ability to have price information dynamically updated via a dropbox account might prove useful for field staff out on a sales call such as for cross-selling. Also, as I mentioned services like dropbox are beginning to offer APIs, so it is only a step away before we end up with new applications that integrate dropbox-based databases into our personal phones and personal lives. These might or might not take off, but the opportunity is there for entrepreneurial firms to explore.

    Habibullah – thanks for your comments. It is very exciting to hear that VMware is close to a solution. I also heard last week that IBM is now able to offer a USB-based solution that boots into a secure area and links to cloud-hosted Windows and Linux desktops: http://www.zdnet.co.uk/news/security-management/2012/03/07/ibm-puts-secure-windows-linux-in-the-cloud-via-usb-40095212/

    The security risks are there, but we are already facing such risks even with the current system, i.e., with existing computer and communications systems. Mitnick and others make a persuasive case that the real risks in security are not just technical but through social engineering as well as insider vulnerabilities.

    Again, thanks for everyone’s comments. I think the next few years will bring exciting changes ahead.

    best regards
    kwang

Leave a Reply

%d bloggers like this: