IBM's cloudy future

ibm bluemixLast week, I attended IBM’s Cloud Summit in New York. This was IBM’s opportunity to re-affirm its commitment to the cloud, and to enlighten us about IBM’s future. Over the course of the day, IBM Cloud’s senior management team detailed their cloud strategy, execution, and future plans.

From the volume of data that IBM presented, here are key takeaways:

Mixing up the brands, bringing Big Blue to the cloud

Bluemix, IBM’s latest rebranding of their cloud business, reports more than $10B revenue, ~ $4.5B for “as a service”. The rest is consulting, outsourced private cloud, and other services. While some are tempted to compare IBM’s revenue and growth (46% year over year) to Amazon, Microsoft, and Google, Robert LeBlanc, SVP Cloud, was emphatic in stating “It’s not a race for size, it’s a race for value!”

IBM’s stated philosophy is to find and obtain innovation by any means necessary, which drives IBM’s recent splurge on acquisitions. On the infrastructure side, Cleversafe gave IBM a local object storage solution that is being integrated into Bluemix for private, hybrid, and public cloud deployments. The Weather Company gave IBM a corpus of data for training Watson, plus the knowledge and ability to handle billions of data points per data, helping IBM to add IoT services to Bluemix. They’ve also invested heavily in video acquisitions, viewing video as the growth driver for the next business cycle.

Watson — come here — I need you

If there was one message that came through loud and clear, it’s that Watson is IBM’s crown jewel — and is the key differentiator for Bluemix.

IBM is building Watson’s machine learning and cognitive capabilities and services industry-by-industry, starting with weather and health, which is, in part, driving IBM’s acquisition strategy. In addition to the Weather Company, the Merge acquisition came with 350 million images, which helped build a corpus of knowledge for both radiology and oncology. Overall, IBM invested heavily in training data for Watson Health, getting 90 million lives worth of data and billions of images.

IBM is also partnering with CognitiveScale and others to increase Watson’s value. CognitiveScale is creating pre-trained Watson powered, industry specific ‘digital brains’ to speed time-to-business-value. CognitiveScale claims it takes 10 minutes to provision your cognitive cloud, 10 days to build your first cognitive business app, and 10 weeks to move the app into production.

Cognitive businesses benefit from putting to work all forms of data, facilitating evidence based, confidence-weighted decisions. This drives insights and pattern recognition with a system that learns and adapts, and navigates natural language to allow conversational interaction.

As a result of these efforts, Watson is showing 100% year over year growth, the fastest of any Bluemix service.

Taking a can opener to Bluemix

Another theme that permeated IBM’s strategy and execution is “choice with consistency”. Simply put, IBM is not picking winners and losers. Instead, IBM is opening up Bluemix, providing a multitude of choices. IBM now believes that clients must be able to choose what works best for them–run time environments, programming languages, deployment models, etc. IBM provides consistency in abstraction layers, API, service, reliability, and business models.

IBM is embracing an open, cloud-first approach, so that clients can take advantage of emerging deployment models, and embed intelligence in the data that sits in their market place and business.

IBM is also warmly embracing the hybrid cloud, recognizing that for many reasons not all apps or data sets will migrate to the public cloud. As companies look at existing operations, the cloud becomes more than just cost savings. You don’t just want to port an application to the cloud, you want to refactor the app, and often the business process and business model.

Cloud clients are less concerned with moving existing apps to the cloud, and more with developing new apps and building the next generation businesses. As Robert LeBlanc said, “the Line of business doesn’t care that you’re in the cloud or in a car. They just care about the value you provide.” Thus about 80% of IBM business is new application development.

Can the biggest of the Big (Blue) learn from the smallest of the small?

Big_and_small.jpgThe Cloud Summit was also the launch of IBM’s 7th Bluemix Garage, in New York. The Garage is IBM’s attempt to formalize Silicon Valley lean startup methodology, agile development, and DevOps into a process for rapid business transformation (The Bluemix Garage Methodology). Going all-in on the Silicon Valley ethos, the garage comes complete with the open bullpen environment, public coffee shop at the entrance, and semi-raw space architecture.

In spite of their desire to foster startups, 70% of their engagements are with existing IBM Enterprise customers, which demonstrates IBM’s strategy of evangelizing new development and new IT models to the conservative, risk-averse traditional business. IBM is also embracing lean and agile in developing the garages — they were working with customers well before the official launch. The NY Garage specializes in Blockchain to meet the needs of IBM’s FinTech customer base in NY, while other garages specialize in cognitive, IoT, or other IBM specialties.

Will Bluemix cloud kill Big Blue data center hardware?

IBM realized that if they don’t have scalability, performance, and functionality at bottom layers of the data center infrastructures, they couldn’t make the upper layers work well. Traditional data centers are built with ‘federated’ products, adding software and support as needed. This works for small scale, but makes for a complex environment that is slow to respond to changes. Each third-party supplied component must go through a long and arduous interoperability test matrix.

IBM developed components, on the other hand, must only be tested to work inside the IBM data center. This leads to rapid development and deployment — another expression of IBM’s internalization of DevOps, Lean, and Agile development.

IBM’s next-generation infrastructure relies on software-defined everything. The base layer is a programmable mesh fabric where each box uses solid-state storage and supports quality of service prioritization. The fabric relies on IBM designed multiport network cards supporting 200Gbps bandwidth in the first generation, providing an overall 100X increase in internal data center bandwidth. The network cards have their own processor, allowing for both out-of-band network management and a true L2 network for the client systems.

Even power supplies get the software-defined treatment, and include their own hybrid battery-backup system. Like a Prius, the battery can be used to boost available power, enabling IBM to design more efficient supplies for steady state rather than peak usage.

Can the Bluemix pivot bring green to IBM?

Robert LeBlanc made an interesting observation: “The notion of being ‘uberized’ is striking fear into the business, the CEO and the boardroom. Organizations are being forced to innovate, and are looking at new business models to avoid being disintermediated and disrupted.”

IBM really takes this concept to heart, having been uberized themselves many times, building new businesses and business models along the way. IBM’s full-throated adoption of the cloud represents their next pivot, from a pipeline product company to an ecosystem company.

While Watson, Bluemix Garage, and other technologies enable IBM to present their ecosystem with significant differentiators, they face significant competition, all of whom are trying desperately to make it a ‘race for value’ rather than a ‘race to the bottom’. IBM’s pivot has been many years in the making, and there’s still a lot more effort and time before the results will be clear.

Insight on trends in converged infrastructure

Topics: Cloud Services & Orchestration