IBM 2019 strategy – running towards the hard things

 

IBM gathered a set of industry analysts together in Milan to discuss its strategy. Having made many announcements earlier in the year at its Edge conference, it chose not to cover much in the way of systems and software news; instead it drew a picture of how it’s helping its customers to leverage the newest technology to compete more effectively in a world becoming ever more ‘digital’. This post is my assessment of IBM’s 2019 strategy.

The meeting gave a comprehensive overview of its approach to:

  • Cloud computing
  • Application modernisation
  • The cognitive enterprise
  • Blockchain applications
  • Cyber resilience

In addition it gave us an overview of four projects its researchers are working on in the labs and introduced us to a number of customers and business partners.

Multi cloud computing

IBM was second only to AWS as a cloud supplier in 2018 (see my Figure) – however it is a very different provider. Despite its acquisition of public cloud supplier SoftLayer in 2013,  IBM’s portfolio includes private and ‘hybrid’ (public and private) offerings. It claims to have 600 private cloud customers around the world, of which one-third are in Europe. Like HPE, Cisco, Oracle, Huawei and almost every other systems supplier, IBM uses ‘multi-cloud’ as a term to cover its capabilities and the ‘journey’ it believes its customers and prospects are undertaking.

Its research suggests that:

  • Comparative spending from 2018 to 2020 will differ strongly for ‘dedicated’ (i.e. public) clouds (a CAGR of 18%), private clouds (15%) and ‘traditional’ IT (-9%), that
  • To date only 20% of enterprise workloads worldwide have been modified for delivery from public clouds, leaving 80% to be assessed, ditched, modernised and/or moved.
  • Investigating 113 of its own customers revealed that 94% of them are using multiple clouds and that 67% use more than one public cloud.

In short it concludes that its customers are living in a hybrid and multi-cloud world. It believes that, while many of them got there by accident, it has the expertise and capabilities to support them to move forward by designing a strategy. Supporting other public clouds is a practical issue for IBM and it is beginning to link its activities to – and in partnership with – AWS and Azure (but not yet Google Cloud Services). To support its hybrid approach it has strong ties to VMware and Red Hat, which it is in the process of acquiring. IBM’s multi-cloud reference customers include Lloyds Bank and BNP Paribas, as well as:

  • Alitalia, whose Roberto Tundo spoke to us about his experience as CIO over the last 3 years. Alitalia’s has used its relationship with IBM to change its IT from ‘physical everything’ 2 years ago to a ‘full hybrid multi-cloud’, which in turn has helped the airline (currently in ‘extraordinary administration, somewhat akin to the US’s Chapter 11 protection) become the most punctual airline in Q1 2019 and increase digital bookings significantly.
  • At the meeting we also heard from Pierantonio Azzalini, CIO of Fincantieri’s – an Italian shipbuilding company, which uses IBM at the heart of its a multi-cloud IT approach. The company builds everything from submarines to mega yachts and oil tankers and has a 10-year full order book and sometimes acquires other (sometimes large) companies. Pierantonio noted that there are some vital security and regulatory applications shipbuilders have to address, but that the move to ‘digital’ means that the pace of change is accelerating. He pointed out that even banks and insurance companies need ‘digital innovation’ – something they didn’t need 10 years ago.There’s an interesting analogy to be drawn between public clouds and servers; users need to make up for the incompatibility of leading services, components, operating systems, programming languages and techniques, etc. through the use of open source components and federated management approaches; a ‘multi cloud’ strategy is similar to the ‘multi cluster’ strategy many have (or are still trying) to master.

IBM cloud offerings can be delivered with EU Cloud Computing Compliance Controls Compliance (C5) accreditation, providing security and data management for multi-cloud applications – useful for GDPR compliance where IBM acts as the data processor.

Application Modernisation

Enterprise applications vary by the server type, its operating system, hypervisor and/or container virtualization in use (see my Figure, which demonstrates how Linux and Docker in particular are common components in an otherwise differentiated landscape). One reason why 80% of all workloads have not yet been moved is that enterprise customer find it a very hard process; so IBM is advising its customers to move, build, manage, govern and/or modernise their applications… and trying to make it a simpler process.

At the heart of the standards IBM uses with its customers is Kubernetes – the open source container orchestration system for automating application management – originally developed by Google, which manages container-based Docker applications. IBM claims there are over 120 Kubernetes versions available to enterprise developers around the world, adopting some of which will leave customers locked-in to their supplier. It believes that its approach is open, allowing customers to choose to use other suppliers, or bring cloud applications and/or data in-house more easily and cheaply.

Fortum is an IBM reference account for application modernisation. At the meeting we heard from Arun Aggural (SVP of Business Technology), who gave an overview of how this electricity wholesaler and retailer is addressing the issue. Its business involves providing electricity from a mix of hydro, nuclear and wind plants, offering spot prices in a market that traditionally works with fixed contracts and struggles to provide adequate supply at times of peak usage.

The company realises it needs to reduce the complexity of its IT operations to become a fully ‘digital’ provider. It chose to use public clouds for running its applications, rather than building its own large data centre; one reason is that, with each plant having between 5 and 10 thousand monitors, it doesn’t want to handle the resultant 1 million points of data a second on its own IT estate. Acquisitions in order to get consumer electric car charging have increased the mix of applications significantly.

Arun started by assessing the applications the companies used first and discovered:

  • The company had 780-800 applications in total and
  • ‘Non-core’ applications were costing 20% of his budget – in one case they discovered one application being used only by a single employee was costing $20k per month.
  • He decided to decommission around a quarter of these applications – a delicate process requiring management support in some cases – and to ‘harmonise’ another quarter. Eventually the company aims to run 80% of its workloads on public clouds, which will be reduced to just Azure and AWS from the multiple suppliers it currently has. Arun remarked that IBM hasn’t (yet) suggested moving everything to its own cloud services. In its electric car charging business he wants to reduce the number of development tool chains from 8 to just 1 or 2. As for application modernisation it has upgraded the operating system and/or databases on 70-80 key applications and is rewriting around 40 from scratch.

The cognitive enterprise

‘Artificial Intelligence’ (AI) and ‘machine learning’ (ML) are generic phrases commonly used by enterprise IT vendors. In practice they are hard to define. Over the decades IBM has chosen to demonstrate potential uses by building systems to compete against humans in chess (Deep Blue beat Garry Kasparov in 1997) and in the TV game show ‘Jeopardy!’ in 2011 (Watson). Since 2014 it as also been developing ‘Debater’ – an AI system designed to present and argue against humans in natural language.

In the conference in Milan it set out 7 keys to success for organisations wanting to become ‘cognitive enterprises’.

 

 

In particular:

  1. Create platforms to unleash transformation,
  2. Leverage the incumbent advantage in data,
  3. Architect the business for change,
  4. Redesign company workflows around AI,
  5. Get agile, change fast and build things,
  6. Reinvent your workforce to ignite talent and
  7. Win with trust and security.

While this process is generic, the prescriptive elements include experimentation, inventive use of the company’s own data, redesign of the computing infrastructure, the introduction of AI and ML leading to a faster adoption of technology to change the roles and employer its workforce. The addition of trust and security are important, especially as the unethical use of AI and associated algorithms is rife among many organisations at a time when they are barely understood, let alone legislated against by governments. For application development IBM suggests a ‘garage’ approach, to which it supply its own experts if needed.

In terms of the IBM users who spoke at the conference:

  • Alitalia told us that it uses machine learning, but thinks of augmented – rather than ‘artificial’ intelligence in the applications it has built. The IBM partnership (which it modified from a contract signed in 2014 to purchase mainframes) has been an important contribution to making the airline more on time and in increasing its online bookings in Q1 by 18% over the 2018 level.
  • Stefan Furnsinn of Yara, a digital farming company, told us that it wants to help farmers and food buyers to be cognitive enterprises – especially in providing information to its wide international network of customers and economists on how to produce food more efficiently. It uses its global partnership with IBM and its Weather Company to build AI applications and precise weather information.

While the concept of ‘cognitive enterprise’ is a much wider issue than the use of blockchain applications, application modernisation or cloud computing, it is a valid target for organisations to aim at, especially for older ones who feel newer ‘born on the Web’ companies have left them behind.

Blockchain applications

IBM was one of the first of many enterprise IT suppliers to pioneer blockchain applications. It is now one of 18 premier members of the Linux Organization’s Hyperledger project, which also has 194 general, 40 associate and 18 academic members. Hyperledger blockchain applications have been growing tremendously over the last few years – an acceptance which has made up somewhat for the early association of BitCoin with criminal activities and the early major failure of Ethereum.

At the conference we were encouraged to participate in a blockchain demo based on BlueMix development platform. We also heard from a number of its customers on their own adoption. In particular:

  • Yara (see above) is using blockchain as it builds business to business applications linking farmers to economists, wholesalers and consumers,
  • IBM has itself introduced the Food Trust, a blockchain application shared between food suppliers and producers. Its current 80 members include Dole, Driscoll’s, Golden State Foods, Kroger, McCormick and Company, McLane Company, Nestlé, Tyson Foods, Unilever and Walmart – the initial participants. The application provides information on the provenance (or food trace-ability) of produce.

Blockchains supply a means of paying for offerings using a shared trust model, useful for when it is difficult or impossible for participants to trust a third party to handle transactions. While blockchains are new, different and have a potential to disrupt many older shared transaction systems, some analysts fear that they are sometimes being introduced when a standard database approach would be cheaper – especially for the smaller supplier.

Cyber resilience

IBM is one of the largest cyber security providers. Before the conference I visited its X Force demonstration data center in London, which is housed in a shipping container and taken to teach customers about the processes they need to adopt when under a cyber attack. I learnt that many, if not all, of the standard IT processes can be compromised, making it difficult to react and recover systems quickly. I believe we have gone a long way beyond the days when organisations could protect themselves adequately by buying appropriate security software and setting up a Security Operations Center (SOC).

In Milan we heard from IBM’s Felicity March, a reformed ethical hacker. She pointed to the discontinuity in many company’s plans between the CSOs who respond to the attack, the IT team responsible for Disaster Recovery (DR), who have to clean and rebuild the systems and the finance/procurement departments, who seldom include the cost of replacement in buying IT systems. A company that is unprepared for a major attack will lose credibility, revenues from weeks or months of not being able to run its business processes. Felicity noted some major problems suffered in the NotPetya and WannaCry attacks and the consequences of cyber attacks on Maersk, the UK’s National Health Service and others.

IBM offers cyber resiliency assessments and believes that organisations should spend more time and money on preparing for cyber attacks, including regular penetration tests, creating better connections between CSOs, IT and procurement functions and plans for recovering systems. In short we should think more about ‘cyber resiliency’ than ‘cyber security’.

How relevant is this strategy?

One of the difficulties of analysing IT suppliers’ strategies is to understand the subtle differences in the meaning of some very widely used, but undefined, industry terms, which is why I liked Alitalia’s distinction between ‘augmented’ and ‘artificial’ intelligence and IBM’s widening of ‘multi’ to include ‘cluster’ as well as cloud. In IBM’s case I was able to get there without too many headaches.

In Milan IBM presented practical solutions to real world problems organisations have in using IT to their advantage. My assessment is:

  • Cloud computing has matured significantly over the last few years and it makes sense for customers to take stock of their mix of on premise and public cloud resources. IBM’s partnerships with Vmware, AWS and Microsoft Azure have been well thought-out and should help users to swap more effectively between on premise and cloud systems, selecting the most appropriate for specific applications.
  • Becoming a cognitive enterprise is a more generic aim – one which IBM itself has been attempting over the last few years. It requires assessment of existing systems, investment in new ones, changes in personnel and will be a continuous process for most, since there will be new technology and business challenges ahead for almost all companies. In talking about it IBM is able to demonstrate the tools (such as Watson for AI and ML) and expertise it can add to help make the changes. I believe IBM is more ethical than most other suppliers in understanding the implications of AI and ML.
  • Application modernisation is a practical task for most organisations, although not all will have as much of a challenge or desire to change as Fortum. IBM’s approach will take a leap forward once its acquisition of Red Hat is completed later in 2019.
  • It makes a lot of sense for users to concentrate on cyber ‘resiliency’, rather than just ‘security’. Being fully prepared for the (almost) inevitable attack by forming teams, testing communications and actions and disaster recovery implications will save organisations from the most severe consequences. A stitch in time saves nine! has been in IBM’s strategy for many years.

IBM continues to prove that it’s keen to innovate and reinvent computing at a time when the vast majority of vendors are happy to live under the umbrella of another’s ecosystem. It chose not to talk in depth about its servers, storage systems and software; rather it presented the challenges its (mostly large or mid-range) customers have in adopting new types of IT and process to survive and thrive in the future. It should do well in an IT and business world in which the pace of change is accelerating.

©ITCandor Limited – unauthorised copying of this content is illegal and will be rigorously defended by us through court action