Red Hat 블로그
The next entry in our series of 2013 technology and IT trend predictions focuses on our perspective for the cloud industry heading into next year. Here’s our take on what’s to come for the cloud.
Security becomes more consumable.
If you pay any attention whatsoever to tech press coverage and IT industry analyst reports, you know that security concerns about “the cloud” (however that term is being used at the moment) consistently top the list of adoption concerns. Even if naïve cloud safe/unsafe arguments have mostly been retired in favor of more subtle discussions, there's still a lot of complexity and uncertainty.
This is partly because the “security” moniker often serves as a sort of shorthand for a variety of compliance, audit, regulatory, legal and governance issues that are often only indirectly related. It's also because, as an industry, we're dealing often with new approaches to computing and delivering application services that don't have clear historical antecedents and established approaches to mitigating associated risk. As a result, dealing with security and associated concerns in the cloud sometimes seem to require true experts in the field, who are almost by definition in fairly short supply.
That's why we're encouraged by the efforts of organizations like the Cloud Security Alliance (CSA), which Red Hat joined back in October. The CSA's mission is to promote the use of best practices for providing security assurance within cloud computing, and to provide education on the uses of cloud computing to help secure additional forms of computing. While the CSA's work benefits everyone, its most important role may be “democratizing” the process of securing and running clouds so that organizations operating and using clouds don't need security rocket scientists on staff. Expect to see tools for more easily and systematically securing clouds gain more attention in 2013.
But data security and privacy remain vexing, and increasingly high-profile, issues.
At one level, protecting against data breaches in the datacenter is a fairly straightforward security problem without many new wrinkles relative to the practices that IT professionals have been following for decades. However, in many respects, we are in a place that's different in kind from times past.
Some of this difference is about connectedness and scale. While security models have been shifting from walled perimeters to defense-in-depth since the early days of the web and e-commerce, cloud-based applications made up of composable services from multiple sources vastly increase potential attack surfaces. It's a vastly more complicated security problem than setting the ports correctly on a firewall.
Perhaps even more problematic, though, is even determining how specific data and data relationships need to be treated and which laws apply. As Dave Einstein noted in a recent post on Forbes: “Adding to the uncertainty is piecemeal evolution of regulations governing privacy and data security, which depend largely on where you live and do business. Europe, Australia and Canada are in the forefront of tackling data protection, while the U.S. lags, leaving a thorny legal landscape for multi-national Internet companies.”
We expect the overall data security and privacy situation to get worse before it gets better. After all, some of the issues date back to before the Internet went mainstream. The issues have just become more visible and more complicated. We've already seen big fines imposed for even relatively minor medical records breaches. Expect to read about more fines in the coming year but only incremental movement ahead on the macro issues around appropriate uses of data.
Bring-Your-Own-Device doubters reach the fifth step: Acceptance.
BYOD is one of the trends that some like to cite as a key cloud security issue given that it takes control away from IT and puts it in the hands of users. More than once we've read an IT professional opining in the comments to an article along the lines of “Just you wait. Enterprise IT departments are going to come to their senses and take the iPads out of those darned kids. And get off my lawn.” (Or something along those lines.)
The thing is that those “darned kids” probably include the CEO and other executives. And look around any organization that's not part of the government or in a highly regulated industry and, chances are, most of the smartphones you see aren't company-issued and provisioned. And the tablets that you probably spy as well are far more commonly purchased by employees for some combination of personal and work use—to the degree that we can even still draw a sharp line between such spheres of activity in general. Bring-your-own-PC is a more complicated issue, for a variety of reasons, but PCs are being “consumerized” as well.
In most cases, BYOD is going to require IT departments to do some combination of rolling out new products, educating users and adopting new processes. At the very least, they need to understand potential exposures and come up with a plan for dealing with them. But just saying “no” isn't a realistic option for the large majority of organizations. And that means acceptance is the only reasonable path forward.
Hybrid shows up in ever more conversations.
IT consumerization is also one component (though only a component) of another cloud computing trend—hybrid cloud computing. Hybrid commonly refers to cloud management that spans both on-premise (or dedicated resources at a hosting provider) and multi-tenant public clouds—although clouds can be heterogeneous in other ways as well.
The consumerization angle is that early public cloud usage was often characterized by users gaining access to computing resources with a credit card because their IT department wasn't moving quickly enough. Such usage can also be outside the scope of any IT governance practices. That can be good for flexibility and speed but it can have a stark downside if there's a data breach or if an application developed using a public cloud can't be easily put into production on-premise.
The idea behind a hybrid cloud is that resources can be made available to users as easily as if they were accessing a public cloud while keeping the process under centralized policy-based IT management, as you can using Red Hat's CloudForms’ open, hybrid cloud management. Organizations are also increasingly looking to hybrid cloud architectures as a way to have a more dynamic computing architecture over time. There are only a modest number of hybrid computing architectures in production today, but the movement towards hybrid is clear. That's why industry analysts such as Gartner are recommending that organizations “design private cloud deployments with interoperability and future hybrid in mind.”1 Expect to hear even more about hybrid clouds in the coming year.
OpenStack demonstrates the power of community innovation.
Openness is one of the most important enablers of hybrid IT because it helps users avoid lock-in to vendors and specific ecosystems. And not just open source but openness across multiple dimensions including APIs, standards and the the requirement that permission to use intellectual property, like copyrights and patents, must be granted in ways that make the technology open and accessible to the user. Openness is also about having vibrant, upstream communities that are at the heart of the innovation that the open source development model makes possible.
The OpenStack Infrastructure-as-a-Service (IaaS) project is a great example of community-driven development. That community is a big part of the reason that Red Hat joined the OpenStack Foundation, of which it's a Platinum Member. Red Hat's a big contributor to OpenStack, the second largest by last count, but it's a broad community with more than 180 contributing companies and 400 contributing developers.
We believe that 2013 is going to see all that developer involvement lead to commercial product in the same way that the open source development model has led to innovative products in operating systems, middleware and countless other areas. The details of how OpenStack is developed and governed have their unique nuances—as is the case with pretty much every major open source project—but ultimately the breadth and depth of community has to be counted as a huge strength.
Private (and hybrid) Platform-as-a-Service (PaaS) goes mainstream.
Like other aspects of cloud computing, PaaS has evolved in response to the market. The basic idea of PaaS—that many application developers don't want to be exposed to and have to deal with the underlying operating system and associated plumbing—remains in place. However, PaaS platforms that limit developers to a specific language on a specific hosting platform have only seen lukewarm acceptance. And it's telling that a number of language- and framework-specific PaaSes have shifted toward a more polyglot (multiple languages/frameworks) strategy.
However, for many organizations, moving all of their development into a public cloud is too big a step even if they can choose their tools. Alternatively, they may simply not want to give up some of the features, such as auto-scaling and application multi-tenancy, that a PaaS can provide once they move an application into production on-premise. The management tools a PaaS like Red Hat's OpenShift offers to system admins can help address these issues and demonstrate that a PaaS needn't be just a tool for developers.
Thus, as has been the case with IaaS, we expect that PaaS is going to increasingly be seen not just as a public cloud capability, but as a private and hybrid one. Perhaps even primarily as private and hybrid, at least as far as enterprise application development is concerned. There are already some early examples of private PaaS in the market but we predict that the trend is going to really accelerate in 2013.
1 Source: Design Your Private Cloud With Hybrid in Mind 24 February 2012 #G00230748