8 Comments

Summary:

HP’s MagCloud is pretty cool, as services go. (Even the New York Times seems to like it.) For only 20 cents per page, it lets small-scale magazine publishers with no use for traditional, large-scale printing services create their own high-quality magazines. The MagCloud site also serves […]

HP’s MagCloud is pretty cool, as services go. (Even the New York Times seems to like it.) For only 20 cents per page, it lets small-scale magazine publishers with no use for traditional, large-scale printing services create their own high-quality magazines. The MagCloud site also serves as a virtual newsstand where HP handles everything: customer buys a copy, HP outsources the ad hoc print job, and printer ships magazine to buyer. All the publisher has to do is create and upload the magazine.

However, while involving the web and a pay-per-use model might put MagCloud “in the cloud” in the most liberal interpretation of the phrase, it is far from a cloud service. HP should leave MagCloud in its printer division and save the cloud talk for something involving, well, an actual cloud. Vendors who stand to benefit from the cloud (HP, for example) are in serious jeopardy of stretching the already-overdone “cloud” label too thin — and MagCloud is a particularly egregious case in point.

Here’s why:

1. There is no computing being done in the cloud. You can’t say something is “in the cloud” if there is no computing being done on a cloud. If there is no unique software involved (other than InDesign or QuarkXPress, maybe), nothing dynamically distributing load across a collection of machines, it cannot be associated with cloud computing. Is MagCloud a service? Yes. Is it a cloud service? No. At this point, it appears to be little more than CafePress for magazines.

2. It’s not multi-tenant, and it can’t scale. In a cloud, one virtualized box can handle the work of numerous users at once. In MagCloud, it appears that one printing press can handle one job at a time. Traditional printing operations are hardly analogous to the cloud, but, like multi-tenant cloud architectures, they at least achieve economies of scale. And scale? If a publication takes off thanks to MagCloud, it’s “Goodbye, MagCloud. Hello, traditional printer.” The price would overwhelm the publisher, the volume would overwhelm the printer, and the sales model would make the MagCloud marketplace obsolete.

3. It requires an expensive, special-purpose piece of hardware on the backend. But the cloud is the anti-mainframe. In cloud environments, you buy commodity hardware. If a box breaks, you simply replace it. If we stretch the definition of a cloud service to cover anything where both the Internet and some sort of service are involved, the necessity of the Indigo press still makes MagCloud unworthy of the label. If a machine in this “cloud” goes down, it takes half a million dollars to replace it — and I’m guessing more than a few minutes. This collection of digital presses is no Google infrastructure.

The IT world has been pretty accommodating to bending definitions of “cloud,” but there has to be a connection beyond just the web. HP is making some legitimate efforts around cloud computing, including its SaaS-y new Cloud Assure, but MagCloud is not one of them.

  1. Hi Derrick,

    Andrew Bolwell from HP here. Wanted to clarify a few things that seem to be misunderstood about the IT underlying MagCloud:

    To your first two points: Sitting behind the MagCloud website is an HP-developed software platform that runs on hosted servers in the ‘cloud’ to handle the processing and management of MagCloud-related compute tasks. This platform needs to be highly scalable to handle millions of user requests and has been developed using a distributed computing architecture to ensure it can meet target performance levels and react to surges in end-user traffic.

    To your third point: The MagCloud software platform also coordinates print requests across a global network of print locations, which serve as print ‘off ramps’ for the magazines that are ordered through the service. Semantically, saying MagCloud shouldn’t be referred to as a ‘cloud’ service because the output is a printed magazine is debatable; however, the exact nature of the IT and press technology is not.

    It’s also worth noting that MagCloud was one of a few ‘cloud services’ featured at Microsoft’s PDC (Professional Developer Conference) last October to showcase their new cloud platform Azure.

    If you’d like to discuss this further, feel free to contact me: andrew dot bolwell at hp dot com

    Best regards,

    — andrew
    Chief MagClouder

    Share
  2. Thank, i am firstly hear of the HP’s MagCloud ,but when try this tools for my vista ultimate,it can’t run now:(

    Share
  3. I’d be interested in Derrick’s comments on Andrew’s points, above. It appears that Derrick posted this without actually talking to HP and finding out how the platform works – Derrick, is that true?

    Share
    1. Derrick Harris Thursday, April 2, 2009

      In regard to Andrew’s comments:

      1. I would posit that any e-commerce platform worth its salt can handle concurrent users, but that doesn’t necessarily make it a cloud. (Although I acknowledge that the project would be a lot more cloudy if it started incorporated the design aspects mentioned in the NYTimes piece.)

      Additionally, the real end-service here is the printing job, so it really is the printing function that needs to scale. You can’t just replicate large, expensive presses on demand. From my perspective, it would be no different than Amazon Web Services having a great, scalable interface, and then offloading the tasks to nonvirtualized, non-commodity boxes. Or mainframes. Maybe it’s the nature of presses that they just don’t scale well. If, theoretically, magazines were to become very popular, they could overtake the available capacity pretty easily it seems (unless HP has a HUGE network of presses working on this). And customers certainly wouldn’t want to pay 20 cents a page at that point. With a cloud, unless security is your concern and you want to buy your own equipment when you have the money, the theory is that you don’t (or at least can’t) outgrow it.

      Amazon Turk and uTest, for example, use humans as resources in the cloud, but humans can be scaled up — and replaced — relativey cheaply and easily.

      2. My main concern is about marketing. In a webinar introducing MagCloud, it is prefaced with a statement about how this will help HP’s printing division, which seems to be where most, if not all, of the revenue will be coming from. There are people who believe SaaS shouldn’t be referred to as “cloud,” or managed hosting is not sufficiently cloud-like to bear the label. The fear is that term loses all meaning if it is applied to everything, and customers get sick of hearing it.

      I happen to think MagCloud is a great idea for a service, I just think it is not a *cloud* service. Obviously, that definition is up for debate.

      Share
  4. Andrw — as you know, I think that what you’re doing with Magcloud is revolutionary. I look forward to the day when its services are available in a distributed way so that other publishing initiatives can use it for print-ordering. We’d love to use it for our http://printcasting.com citizen-publishing initiative in that way.

    Dan Pacheco, Founder of Printcasting.com

    Share
  5. Whether MagCloud uses “cloud” approaches or not is an interesting question. But we should first define what “cloud” approaches are. If I listen to IBM, everything is “Cloud”, so, MagCloud being part of everything, it is the cloud also.On top of that, your comment that presses are not virtualized etc. is a little funny. If I take the definition for cloud that it’s a service, available on demand (provisioning in a couple minutes), payable by credit card and that can scale, the MagCloud is a cloud service as it addresses all these. There is nowhere written that cloud services cannot have a physical component in them for example. So, I fully agree there needs to be a clear definition of “cloud services”, but let’s take the right target.

    Share
  6. Lol. To most ppl this is article is full of Tech jargon and are like, what? Is it gonna rain up in here or what? Only you IT geeks truly care about how “cloudy” MagCloud is or isn’t and what not. I mean there’s nothing wrong with that. It’s your job to care… but not ours. The Publishers, Editors, Journalists, and even a school’s PTA benefit from MagCloud and just want it to work. As far as I can tell it’s a pretty well-oiled machine. Thanks HP! :D

    Share
  7. [...] HP lancé en 2008, qui imprime ses magazines à 20 centimes par page. Même si MagCloud a été fortement critiqué par le blog américain GigaOm pour son usage inapproprié du mot “cloud” l’année dernière, la plateforme [...]

    Share

Comments have been disabled for this post