IBM and Red Hat Take on the Edge

The carrier cloud and edge computing stories just never seem to end.  Just yesterday, I noted the trend in edge computing that promised a bit of action in the edge, relating to 5G and carrier cloud.  It looked like public cloud players might be seizing the 5G edge.  Now, IBM and Red Hat have thrown their hat into the same ring.  The picture they paint is very similar to the one Microsoft painted with its own Azure Edge Zone story, and the issues they face are similar to those I cited in yesterday’s blog.  The difference is that IBM and Red Hat have jumped all the way to that generalized edge model that I talked about yesterday, and that might create a very interesting market dynamic.

The key line in the announcement is that IBM and Red Hat “announced new services and solutions backed by a broad ecosystem of partners to help enterprises and telecommunications companies speed their transition to edge computing in the 5G era.”  Microsoft seems to have taken their shot at the edge by example, by demonstrating the application of Azure to operators’ 5G virtual infrastructure mission.  IBM/Red Hat are shooting for an architecture, an edge computing model that enterprises can adopt too, and viewing 5G as a broader driver of edge computing.

The key technical point in the announcement is that containers are the vehicles that will take applications to the edge.  IBM links their story to Red Hat OpenShift, which is a container and Kubernetes ecosystem.  That effectively makes edge computing a resource in a container world, and unifies the edge, the cloud, and the data center.

One reason this is important is that it solidifies the view held by many (including me) that containers are going to displace virtual machines in nearly all applications.  You can still do VMs with OpenShift, of course, but enterprise IT is going to containerize so completely that those who don’t use containers and Kubernetes are going to be at a growing strategic and platform software tool disadvantage going forward.  If you’re not Kubernetes, you’d better get there.

The other reason this is important is that it shows IBM believes that edge computing will not be dominated by specialized appliances, but will instead be an extension of the cloud.  The specialized edge model, including the use of real-time operating systems like the RTOS approach Microsoft favors, would be expected to run a static application set, not one requiring container/Kubernetes dynamism.

This is an enterprise perspective.  For the telecommunications companies, specialized open devices seem to be favored.  However, while you can host specialized static applications on generalized cloud resources, you can’t run generalized applications on specialized appliances.  IBM is betting the edge market will converge, not fragment into mission-specific approaches.

This doesn’t mean IBM is abandoning the telco space, of course.  In fact, they announced the “IBM Telco Network Cloud Manager – a new solution offered by IBM that runs on Red Hat OpenShift to deliver intelligent automation capabilities to orchestrate virtual and container network functions in minutes.”  This seems to blow politically motivated kisses at the telco virtualization specifications while delivering virtual function hosting in what’s a general cloud-and-container model.

IBM is also addressing the IoT space, something that the telecommunications companies may eventually get to.  Their IBM Edge Application Manager is designed to support the management of up to ten thousand edge nodes by a single administrator (for basic pattern and policy-based commands).  This seems to be a refactoring of the basic OpenShift operations tools, aimed at offering businesses with large-scale IoT plans the confidence that operations complexity won’t bury their staff.

As far as applications go, IBM has two layers of support.  Basic edge platform tools, including IBM Visual Insights, IBM Production Optimization, IBM Connected Manufacturing, IBM Asset Optimization, IBM Maximo Worker Insights and IBM Visual Inspector, are available to augment the container deployment framework.  In addition, they announced the IBM Edge Ecosystem, which is a partner program to create vertical and horizontal solution frameworks and promote them to buyers.

There are some semantic points here too, that even though they seem terminological in nature are quite important from a market perspective.  The first is that you’ll note that the announced products and programs are IBM-branded rather than branded or co-branded by Red Hat.  Some might see this as simply a reflection of the fact that IBM bought Red Hat, but I think it’s more fundamental.  Edge computing is a big-player opportunity, and IBM has special strategic credibility and influence in the big player space.

IBM is the only vendor to have ever achieved dominance in strategic influence over enterprise IT spending plans.  They don’t have the same level of influence today as they had at their peak, but they’re still a powerhouse among the larger enterprises, utilities, and financial players.  These spaces have strong potential for edge computing, so having IBM take the lead makes sense.

The next point is that the announcement names Vodafone Business, Samsung, and Equinix as early adopters, and provides a brief summary of the edge offering(s) each has begun to deploy.  The applications demonstrate the edge and hybrid cloud, 5G exploitation, and remote work.  None of these are the host-my-virtual-functions-for-5G, which again demonstrates that what IBM is doing is defining an edge architecture not building a one-off edge package.

This may be the smartest thing IBM could have done, overall, even if IBM in the long run wants to take a big bite out of the 100,000 data center’s worth of work that carrier cloud represents.  5G virtualization infrastructure is just one driver of carrier cloud, a driver that’s probably not worth chasing if hosting virtual functions is the only business opportunity.  IBM isn’t even betting that 5G will have broad systemic impact in things like IoT.  What they’re betting on is a form of edge computing where the edge is just the scouting force of hosting resources pushed out toward the IT frontier.

They may be reflecting the historically dominant trend in IT, the thing that’s advanced IT and IT spending since the 1950s.  Getting processing closer to the worker, providing IT at the point of activity, is the mission of any IT giant who wants to ignite a new wave of productivity-driven server and software spending.  If IBM could gain an upper hand in edge computing as a means of delivering point-of-activity worker empowerment, they could see a major uptick in revenue, enough to free IT spending from the more-for-less pressure of the last two decades.

And keep this in mind; none of IBM’s example applications were 5G infrastructure.  We have to wonder whether this is simply backlash against Microsoft’s Azure Edge Zones announcement, or whether IBM thinks Microsoft has made a rare strategic blunder, one that could give IBM a chance, not only with edge software but for IBM’s own cloud.  I don’t think IBM is writing off network operator 5G deployments, but I don’t think they’re depending on them.  That might be a very wise move.

The question now, clearly, is whether IBM’s and Red Hat’s competitors will make similar moves.  VMware, Dell, and HPE have similar aspirations, and Amazon and Google could take a different stance on edge computing than either IBM or Microsoft.  There’s even more reason than usual to think that the edge space could benefit from some competitive jockeying and creative alternative business and technology models.  If that happens, we could see a lot of dynamism in a space that so far has been dominated by hype.