Big Blue is buying Red Hat for $34B. The deal should close in the second half of 2019.
Is “Big Purple” in your future? Well, that depends on what you think are the answers to these questions:
- By some estimates, the volume of data increases 10x every two years. How fast is the volume of software code – which creates all that data – growing?
- A significant portion of that data will be generated at the edge. What percentage of software yet to be written will be written for the edge?
- How will edge solutions (including digital twins) be crafted and distributed? As highly proprietary commercial software in which the software vendor essentially “steals” client intellectual property and sells it to others using high-priced software licensing schemes? Or through an open source mechanism that supports sharing while sustaining IP protection and keeping aggregate software costs across trillions of devices manageable?
- Most of the edge (and much of the data center core) can’t move to the cloud. Instead, the cloud will have to come to the edge (and core), in hybrid form. Can enterprises practically construct hybrid cloud out of dozens of environments or are they going to have to focus on one or two? Will enterprises use a private cloud approach to evolve their high-value legacy applications?
- Will enterprises lean heavily on service companies to help build, package, deploy, and service highly distributed, high-value, and cloud-based edge and core systems?
- Will those services be a combination of consultancy-based and cloud-based at first, but moving more to cloud-based over time, thus making them potentially more lucrative for “hybrid” service companies?
- Do enterprise users and key channel partners trust IBM and Red Hat?
- Is this deal good for both IBM and Red Hat? Is it a good deal for enterprise customers? Partners?
That’s a lot of questions and I don’t have precise answers for all of them, but let’s jump through them quickly and then get to the answer for (8).
Let’s clump the answers for (1), (2), and (3) – which consider the importance of open source in the future – together. The world is going to need more software than ever and can only acquire it through an open source model. Why? Because the lowest cost and highest quality approach to software development is reuse combined with a commons-based licensing model. The population of software developers may be growing by 15% per year and improvements to the quality of tools may be improving developer productivity significantly, but there is no way the specificity and complexity of edge and increasingly integrated core solutions can be developed in a few software houses or using a bespoke soup-to-nuts approach in enterprises.
Will we see better “edge-to-core” frameworks for building software? Absolutely. The Kubernetes ecosystem, for example, has that problem in its sights. Can we add IP protections to the basis of open source to allow enterprises to appropriate returns to edge-related invention? Blockchain is an example of a technology that could be used to track and manage contribution of code to the open source commons. How will we price software for trillions of devices? Well, let’s claim that the “real world” to “digital world” software interface is going to catalyze the greatest jump in code production in history, and the vast majority of it will be based on open source.
What about questions (4), (5), and (6), which focus on how cloud-based solutions will be written and operated (given that open source software will be the greatest contributor to those solutions).
Wikibon strongly believes that the default approach to cloud will not be “move all data to the cloud,” but rather “move the cloud experience to the data, both at the edge and in data center cores.” Why? Physics (e.g., latency and data locality), economics (e.g., bandwidth costs and IP protection), and regulations (e.g., privacy and governance). This guarantees that hybrid cloud will be the dominant framework for enterprise cloud architectures.
But the enterprise goal should be to distribute the data and processing to where it’s required, while limiting the complexity of the hybrid cloud platform employed. The likely scenario is that enterprises choose a relatively simple, but powerful hybrid cloud plane that is plastic (e.g., scales and easily reconfigures in response to workload changes), can be managed from within the business’s asset portfolio, is highly secure, features predictable costing, and can easily acquire value-adding cloud services at the level of SaaS and PaaS. Simply put, the more common – and open – the platform across cloud, core, and edge, the better for the enterprise.
However, a simpler platform doesn’t predict simpler applications. The solutions being built to traverse the real and digital worlds require unprecedented combinations of deep technical (i.e., evolving cloud, security, and AI technology), physical (i.e., real-world events being tracked and acted upon), and business (i.e., emergent and contingent processes and agency regimes) expertise. These solutions certainly will be applied using significant service resources, both of the professional service and cloud service kind. Why? Because expertise will accrete fastest in professional service companies and computing scale in cloud service companies.
The big trick for service providers, of course, will be how fast technical, physical, and business expertise can be translated into software. As expertise is coded, the relationship between service types will shift from lower margin professional services to higher margin cloud services. Which will fuel more technology, physical, and business research and development, which will turn into more software, etc. Of course, as long as that software mostly follows an open source lifecycle. Otherwise, the wheels of progress will gum up.
So, that suggests that combining a company like Red Hat with an extensive open source presence and a company like IBM with extensive core-to-edge technology and business expertise should be a winner, right?
Well, the devil is in the details and my colleague David Vellante has posted an extensive review of those details. His conclusion? Probably, but we’ll see.
That leads to the answers for (7) and (8). I think I’ve shown that the vast majority of software hasn’t been written, that an open source approach (like Red Hat’s) will be essential for it to be generated and adopted, and that the solutions to be built will create unprecedented and fungible opportunities for turning expertise into code (which IBM does well), much of which will run at the edge and in core systems on-premises (which sustains the relevance IBM’s legacy customer relationships).
If the market trusts IBM and Red Hat (and IBM doesn’t abuse that trust) then it’s a good deal for users and partners. Certainly, we have to wait and see, but at Wikibon we’ll focus on a few questions over the course of the next year as we wait for the deal to close, including:
- Will IBM and Red Hat together create “de facto” solution targets that the open source world excitedly embraces and builds to? One of the reasons big data underperformed is that hundreds of commercial contributors focused on building a zoo of open source software and service options instead of clear business outcomes. For this deal to achieve its potential, IBM must become a business solution lynchpin in open source. Today, Google does this better than any other company. IBM is just one of many.
- Does IBM embed Watson services in open source software utilizing a credible “open source-friendly” pricing and packaging model? Watson is underperforming in the market, and pricing and packaging are key reasons. It’s not being embedded in third-party (or even IBM) solutions as fast as it needs to be. If Watson remains positioned as a business luxury instead of a business commodity, then IBM will continue to be torn in multiple directions and ultimately unable to lead the market in translating business and research expertise into software.
- Do IBM’s legacy customers start to move to the cloud in 2019? One of the key drivers of hybrid cloud will be high-value legacy applications that can’t easily move to public cloud. How fast IBM can bring the cloud experience to these enterprises in an integrated, holistic way will be a crucial question for their ongoing customer relationships. Some, like American Airlines, are committed to IBM’s cloud platform, but list appears short. IBM faces serious technology, business model, and pricing questions here, but the need to show a return on their Red Hat investment should help.
- What about IBM’s competitive, multi-vendor software portfolio? IBM has been delivering more competitive software of late, specifically in the storage domain. Will IBM’s multi-vendor, but still proprietary, portfolio be enhanced with Red Hat’s portfolio? Kept distinct? Starved for investment? Clues about IBM’s direction here will come early and clearly.
