Monday, December 19, 2016

IBM Systems - year-end review

By Rich Ptak

It’s been a busy year for IBM, what with transitioning, proliferating use cases for Cognitive Solutions, the rapid buildup in their Cloud platform ecosystem and announcements of a series of innovative industry-specific solutions and projects, LinuxONE mainframe activities, the OpenCAPI initiative, etc. You might even expect they might slow down a bit.

Nothing even resembling that appears to be in the cards. IBM CEO Ginny Rometty led off the week announcing that IBM will hire 25000 new US employees while investing $1 billion for employee training and development. This is to take place over the next 4 years.  And, she promises the focus will be on “new collar” hires, i.e. skills-based, rather than focusing simply on “higher education” requirements. The needed skills include cloud computing technicians and specialists in service delivery acquired in vocational training or “on-the-job”. IBM collaborated in the curriculum design and implementation of schools to do the training. You can learn more about these Pathways in Technology Early College High Schools (P_TECH)  here[1].  This spending is in addition to the $6 billion spent annually on R&D projects. This is hardly the behavior of a company unsure about its future. But, I digress.

In the same week, Tom Rosamilia, Senior Vice President, IBM Systems held his annual review on the status of IBM Systems. Taking place prior to year-end results reporting, it was light on financial details. But, the broad strokes presented a company that was seeing both progress and positive returns.  In a turbulent and rapidly evolving business environment, IBM had embarked upon a bet-the-business strategy of transformation and innovation in technology, business models and skills to address the challenges of the evolving era of cognitive computing.

IBM committed to the cognitive era before it was fully defined and clearly defined as a viable market. Early hallmarks of the changing environment were the growth of cloud-based and Infrastructure-as -a-Service enterprise computing. Rosamilia quoted industry research that states that by 2022, an estimated 40% of all compute capacity would be provided through service providers. And, some 70% would be in hybrid (combination of on- and off-premise infrastructure).  Such configurations meant IBM’s systems infrastructure-based revenue streams would need altered delivery and service models to grow. 

In response, IBM Systems altered its operating model to focus on three main areas:
  1. Cognitive Solutions – through partnerships and initiatives that grow the ecosystem and increase use cases, scaling up performance and making it easier to leverage the technology.
  2. Cloud platform – building out the ecosystem, increasing accessibility by building out available services and tools and expanding utility with easy access to the newest technologies.
  3. Industry Solutions – focus on providing servers, services, platforms along with innovative solutions optimized and targeted to resolve industry-specific challenges.

In 2016, the systems product portfolio was divided into 3 areas: 
  1. The Power chip-based systems targeting the market with Open Computing (e.g. Open Compute Project), Power LC servers with NVIDIA NVLink™ for specific workloads models, the OpenCAPI[2] Consortium (dedicated to developing a standard that allows cross-vendor access to CAPI acceleration) and the OpenCAPI Consortium (developing a standard that allows cross-vendor access to CAPI acceleration), and PowerAI –  an IBM partnership with Nvidia that enables simplified access and installation of platforms for deep learning and HPC efforts;; 
  2. the zSystems with LinuxONE for hybrid clouds, high security z13s with encryption for hybrid clouds, Apache Spark support on z/OS and (the blockbuster) secure blockchain services on LinuxONE available via Bluemix or on-premise; and finally,
  3. IBM lit up the Storage /SDI (software defined infrastructure) markets with all-flash arrays available across their complete portfolio and a complete suite of software-defined storage solutions. There are plenty more coming with IBM Cloud Object Storage solution, cloud support for IBM Spectrum Virtualize and DeepFlash ESS. We don’t cover the areas, so we won’t comment more.

IBM will continue to stir things up as they expand and enhance deliverables in these areas in 2017. There is a special focus in Cognitive Computing where speed in data access and computational power are critical. Power systems with CAPI are specifically designed for Hi-speed, computationally dense computing. They benefit from partner developed accelerators. Cost is a major issue in high-performance analytics and computing. Power systems with CAPI and accelerators offer significant price/performance advantages over general purpose systems.

Block chain has been a major news item this past year. Use cases are proliferating as understanding of the technology and accessibility grows. Both significantly benefiting from IBM offerings of easy flexible access options to the technology, as well as training including some which are free.  Financial, healthcare[3] and business use cases for blockchain in secure networks are proliferating.  IBM is and will continue to be a major booster and contributor in the spread of this technology. IBM is offering secure blockchain cloud services on-premise or via Bluemix cloud with either LinuxONE or Power LC systems. 

Rosamilia discussed a number of activities underway with clients and customers expected to deliver in 2017. These include collaboration with Google and Rackspace using a Zaius server (running on yet-to-be generally available) Power9 chip and OpenCAPI to deliver 10x faster server performance on an “advanced, accelerated computing platform…(delivering an)…open platform, open interfaces, open compute”.  There is the blockchain/(HSBN) high security business network with existing and potential application across a range of business functions. These include securities transactions, supply chain, retail banking, syndicated loans, digital property management, etc.
Tom Rosamilia described how Walmart uses blockchain to guarantee the integrity of food from farm to consumer. Sensors packaged with farm products are tracked from farm to final consumer purchase to assure environmental conditions (e.g. temperature exposure) have been maintained. More examples are available (see our recent blog[4] on blockchain).

The session concluded with the list of technologies where IBM is investing. These include POWER9, LinuxONE, zNext (the next generation mainframe!), all-flash systems, next generation computing (beyond silicon), open ecosystems, blockchain and (presumably LARGE) object storage. We believe that their bets were well-placed. There is still much to be done, but it appears to us that Ginny Rometty and IBM will keep every one of those 25,000 new hires very, very busy.

[2] CAPI (Coherent Accelerator Processor Interface) developed by IBM for and initially only available on Power systems. It resolves the increasingly serious data access and transfer bottlenecks problems to improve system performance by a power of 10.)