The Logistics of Information


Published on

Logistics, Data in Motion and Paradigm Shift of the CIO: The economics and psychology of the flow of information. Advances in IT, especially cloud technologies, are causing a shift in the role of the CIO.

Published in: Business, Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

The Logistics of Information

  1. 1. Logistics, Data in Motion andParadigm Shift of the CIOThe economics and psychology of the flow ofinformationVince KellenChief Information OfficerUniversity of Kentuckyvkellen@uky.edu4/29/2012 1
  2. 2. ChangeThe only thing that seems stable is theoverused truism that change is permanent.What is also not in dispute is that the rate ofchange has accelerated and is continuing toincrease.Information technology is the primaryaccelerant in the recent increased rate ofchange.The effects of the increase in the rate ofchange are being felt in all aspects of life:personal, careers, social structures,governments, climate.The increase in the rate of change is withoutprecedent in human history. We are allpioneers. 2
  3. 3. The [gradual, sudden] shift CIO Past • Insourcing networks, data centers and software engineering • Configuration management • Stability, reliability CIO Current • In/outsourcing networks, data centers, enterprise software • Innovation • Transformation CIO Future • Cloud services • Venture investor, organizational psychologist • Information logistics expert 3
  4. 4. What is creating this dynamism? The computer! 4
  5. 5. What is creating this dynamism? Rising information intensity and flow • Massive growth in density, quantity and diversity • Dramatic improvements in breadth, ease and speed of access • Global collaboration Faster innovation, mimicry • Business process replication • IT process replication (e.g., cloud, ITIL, enterprise architecture, PMO) • Speed to market of new offerings IT is affecting industry competitiveness • Lower barriers to entry, increased market turbulence, major competitors changing positions – From “Scale without mass: business process replication and industry dynamics.” E. Brynjolfsson, A. McAfee, M. Sorell, F. Zhu. Harvard Business School Technology & Operations Mgt. Unit Research Paper No. 07-016 (2007). • Individual and organizational human systems are responding to this pressure 5
  6. 6. What are the consequences of this change? Investments in IT capital are continuing • Buoyed by dramatic price/performance gains (e.g., Moore’s Law) First mover advantage might not mean much • New startups grow and die quickly or get eaten by bigger fish • This results in increased consolidation in IT intensive industries • But… remaining dominant players leapfrog each other • And... Industry consolidation grows significantly – Is higher education about to become IT intensive like service and manufacturing sectors? IT leaders will need more financial, strategic and human skills • Continued growth in IT investments demands improvements in productivity (reduced costs, more output) • Market dynamics require nimble architecture that all can enjoy (no sustainable competitive advantage) • The impediments to superior use of tools (in this case IT) will lie more and more in organizational and individual tailoring and adoption 6
  7. 7. Can this changeProbability of occurrence persist? 1 Zone of optimal information advantage (edge of chaos) To gain an edge, To avoid losing, agents will develop agents will develop more complex simpler models, models, with longer with shorter time time horizons, horizons resulting resulting in in decreasing increasing dynamic complexity dynamic complexity 0 Level of Complexity 0 1 Order Chaos 7
  8. 8. The consequences for failing to catch up  Abundance of information Destabilizing without improvements in human gap absorption and decision making 1970? will result in a growing gap in the breadth and quality of information actually used by people and organizations  This gap (a form of information ambiguity) could increase overall instability and uncertainty. Agents operate out of distinct world-views that filter information  This gap makes available competitive opportunity for those who can understand more, provided some balance between order and chaos remains 8
  9. 9. What can slow down dynamism? Human attention • Our ability to process information limited to waking hours • We will need more automated decision making, reserving human attention for strategic maneuvers and change Culture • Cultures and their attendant legal systems are slower to adjust to automated decision making and new concepts Personal and organizational defensiveness • While change brings opportunities, it threatens people • These changes will not be without a battle of sorts 9
  10. 10. Speaking of change, now let’s talk about thecloud… First some IT facts of life 10
  11. 11. ServerServerHuggerServerHuggerTrainee 11
  12. 12. What is this about cloud? The new outsourcing • Cloud represents a new way of integrating technologies (and business processes) so that the institution relies on external vendors for basic services • Cloud is very real, very big and will transform IT • Morgan Stanley May 2011 analysis expects adoption to be about 51% of organizations and about 22% of the IT workloads run in the cloud in three years. On premise growth in servers is expected to be flat or shrink • All major vendors are committing >$1 billion each in cloud technology What makes cloud computing unique? • Widely used, well understood and generic components • Quick provisioning and de-provisioning • Flexible contracting and procurement 12
  13. 13. Cloud vernacular Software as a service (SaaS) • Software hosted elsewhere. Higher education has been steadily adopting SaaS • Examples: Hobson’s CRM, Digital Measures, ServiceNow IT support Infrastructure as a service (IaaS) • Infrastructure hosted elsewhere. Higher education has NOT yet adopted this technology. General purpose server computing can be hosted with a vendor or consortium • Amazon’s elastic computing and storage solutions are examples of ‘generic’ cloud • Large vendors are bringing custom, enterprise cloud solutions forward now Platform as a service (PaaS) • This includes tools to create applications in the cloud • Examples include Microsoft Azure, High performance computing (HPC) as a service may be coming • National labs have long since been an ‘outsourced’ provider of HPC • Expect more HPC university consortiums, offerings be large vendors • Cost of electricity, generic workloads make HPC as a service attractive 13
  14. 14. So, what does this mean for data centers? First, let’s look at an institution’s data center of the present… 14
  15. 15. 15
  16. 16. Now let’s look at an institution’s data center of the future… 16
  17. 17. 17
  18. 18. What does this future look like to our server hugger? 18
  19. 19. 19
  20. 20. 20
  21. 21. Sizing capacityInstead of sizing on-premise computing at 100% of potential usage, size at someamount lower and ‘burst’ out to cloud providers for ‘overflow.’ This can save costs100% ??Utilization 0% Jan Apr July Oct Dec 21
  22. 22. Types of workloads Virtual desktops for student access to academic software • Peak usage in midterm and final weeks • Low usage during breaks ERP processes • Peak usage during enrollment periods • Testing and development of new ERP functionality Storage • Take very large amounts of data, infrequently accessed to large-scale, low-cost providers Disaster recovery • Instead of maintaining a hot/warm site, secure contracts for quickly expanded processing 22
  23. 23. All forms of cloud will be useful Software as a service will continue to be important • Enforce real-time data integration for quick user account provisioning and same-day data analysis • Review contracts for legal gotchas and security holes Infrastructure and platform as a service will become the center of attention • Select multiple vendors to encourage both diversity of supply and competition • Match workload characteristics to vendor strengths Key challenges for the immediate future • Strong ‘cloud orchestration’ tools to help IT manage on/off premise computing with multiple vendors • Easily managed, real-time data integration across multiple cloud providers • Flexible contracting and pricing models, especially in the area of software licensing 23
  24. 24. Examples In use today @ UK • Google and Microsoft email (SaaS) • Digital Measures (SaaS) • Hobsons CRM (SaaS) • Xytracs – Accreditation software (SaaS) Testing today @ UK • IaaS for ERP • IaaS for HPC In use tomorrow? • Business intelligence and analytics in the cloud? • Learning management systems in the cloud? • Supercomputing in the cloud? • Desktops in the cloud? • e-Textbooks in the cloud? 24
  25. 25. Will everything be in the cloud? Not for the foreseeable future (4-8 years) • Costs. We can save costs by running dedicated workloads locally (this may change!) • Risk. We need some ability to be immune to vendor supply chain disruption (other industries get by without any local production!) or the cloud market is emerging and still not mature • Cost of data computing versus cost of data transmission. Some workloads will just be cheaper to process locally than transmit the data across a network (think very large files, super rich medical imaging) • Some software isn’t technically capable of running in the cloud (True!) • Some software vendors can’t figure out how to price their software in the cloud (True!) For all these reasons, a hybrid cloud strategy is important 25
  26. 26. Current model: Slow, big central data center SaaS External Solution A Data Center A Firewall Data SaaS Solution B Center External Data Center B SaaS Solution B 26
  27. 27. New model: Fast, small central data center SaaS External IaaS Vendor B Solution A Data Center A External Data Data IaaS Vendor A Center Center B Integration orchestration and security layer SaaS SaaS Solution B PaaS Vendor A Solution C 27
  28. 28. Moving from ‘anticipate slowly’ to ‘react quickly’ Since the beginning, IT has planned capacity in either 30 year chunks or 5-10 year chunks • 30 years for the design and use of a data center • 5-10 years for the selection and use of servers and storage • Moore’s Law and incremental capacity increases provide the rest In a ‘pull’ approach, IT services are provisioned much more quickly. Using self-service cloud approaches both on- and off-premise this can be in minutes • No lead time in provisioning, no ‘shelf time’ as equipment sits in storage rooms, automated provisioning – Virtual servers, operating systems, application installation, database loading, network provisioning, security policies If the entire system moves faster with far fewer manual steps, what happens? 28
  29. 29. What business benefit does ‘Pull’ really create? Organizations can turn a new idea or change into action much quicker • This can allow for faster delivery and it can allow for faster failing! • Can organizational decision making keep up? • To borrow from True Lean, eliminating waste (time) can save effort/cost, leading to best-in-class efficiency IT shifts its focus from mundane provisioning and support activities to: • Assisting in the design of ideas • Managing suppliers • Optimizing IT costs • Review of the effectiveness of ideas implemented Does this mean IT staff will no longer be needed? • No! • We have more demand for new work than we can keep up with. We can shift people to new roles that are needed to contribute to new work 29
  30. 30. What should this mean for university operations? IT should be directed to: • Measurably, over time, lower the cost of IT relative to revenue • Enable improved outcomes for both learning and research • Accelerate the development of business insight for both cost savings & growth IT should be a scalable infrastructure to help the institution find reallocations and new revenue while maintaining quality Revenue $ Direct labor $ Administrative processes and IT/process infrastructure $ Year 0 Year 1 Year 2 Year 3 Year 4 … 30
  31. 31. (Pause)How can we harness the forces of faster information flow and increased rate of change without losing our minds?
  32. 32. Let’s review the CIO role in this… What do we have so far? • Dynamism (internally and externally) is growing • Business process (and those within higher ed) are replicating quickly • IT processes are replicating within the vendor community and between organizations • The impediments to successful use are likely to be increasingly human, not technical • The investments in this whole ball of wax continues to grow, possibly requiring more complex financial acumen to manage The role of the CIO is shifting • Away from mainly technical infrastructure configuration management • To the dynamics of information costs (technical) and uses (human) 32
  33. 33. The role of organizational capitalInvestments in computers + peopleare synergisticOrganizational capital: • Degree of self-managed teams • Employee involvement in groups • Diversity of job responsibilities • Who determines pace of work • Who determines method of work • Degree of team building • Workers promoted for teamwork • Off-the-job training • Degree of screening new employees for educationFrom “Intangible Assets: Computers and Organizational Capital,”E. Brynjolfsson, L. Hitt, S. Yang. Center for eBusiness @ MIT,MIT Sloan School of Management. Paper 138. (2002). 33
  34. 34. A lot stands between bits and appropriate action The purpose of IT is to improve agent action Action (individuals & organizations) • Both efficiency and appropriateness of action Organizational defensiveness When millions are invested Peopleware and the payoff [is, is not] not Governance, teamwork achieved, who should take [credit, responsibility]? Individual motivation • Can the CIO credibly say Hard/software “They made me do it?” Visualization, usability Infrastructure, ERP 34
  35. 35. Information immunity “You don’t understand. On the 22nd floor, we work in a fact-free environment.” While computing excels in turning data into information, getting information shared with fidelity across human beings is difficult • Different disciplines create different mental models. Information mutates or is rejected, depending on the mental models one has • Double-loop learning is not so common • Individual information processing orientation matters – Are you predominantly a Data->theory or theory->data person? • Power relationships, threats to position, conflict, group-think can filter out information Is this why investments in organizational capital are synergistic with investments in IT? 35
  36. 36. Boyd’s OODA “Loop” Sketch Observe Orient Decide Act Implicit Implicit Guidance GuidanceUnfolding & Control Cultural & ControlCircumstances Traditions Genetic Heritage Analyses & Feed Observations Forward Synthesis Feed Decision Feed Action Forward Forward (Hypothesis) (Test) New Information Previous Outside Experience Unfolding Information Interaction With Unfolding Environment Interaction Feedback With Environment Feedback Note how orientation shapes observation, shapes decision, shapes action, and in turn is shaped by the feedback and other phenomena coming into our sensing or observing window. Also note how the entire “loop” (not just orientation) is an ongoing many-sided implicit cross-referencing process of projection, empathy, correlation, and rejection. From “The Essence of Winning and Losing,” John R. Boyd, January 1996.,
  37. 37. CIO Past Human Market Stable Dynamic Scope Technical Focus of work 37
  38. 38. CIO Present Human Market Stable Dynamic Scope Technical Focus of work 38
  39. 39. CIO Future? Human Market Stable Dynamic Scope Technical Focus of work 39
  40. 40. A future state – financing and contracting Computing supply chain management • Standard, spot, futures, options contracts • Options exchange • Use of brokerage and risk management services • National and international network architecture • Monitoring, measuring, pricing the flow of information between systems • The institution can be both a buyer and seller • Not all architectures will be equal, some snake-oil will be sold The physical logistics of information will become more complex as price/performance optimization options grow • The network will be the critical scarce resource. It is physically and legally constrained in ways all the other layers of IT are not • Data center automation – Scheduler, rule-based, statistical inference/correlation engine • Cost of compute, transport, storage and policy must be determined • Costs + workflow type (CPU, I/O, disk, priority, time-to-market) must be optimally matched to provider 40
  41. 41. A future state – from push to pull Orchestration of pull business architectures • How do you get business processes and technology to be created/altered quickly? • How do to streamline business processes across the board? • How do you effectively and efficiently tailor valuable, core processes? – Research, teaching, community service, healthcare Design of a dynamic hybrid infrastructure • How do you ensure your network will take your data to where it needs to go? • How do you manage security in this more complex network environment? • How do you know when/where to process workloads? – (Compute costs + network costs + policy costs) must be [dynamically, statically] compared with provider capabilities 41
  42. 42. A future state – organizational development “Peopleware” • Once you have a dynamic, cost-effective, economically tractable IT infrastructure, you have to have an organization capable of using it – The Tale of the New Lamboughini • How does the CIO get “footing” to encourage a more rigorous discussion of larger organizational issues? – To the degree that IT is both competitive and costly! Not all organizations will require this full-throated conversation (yet) • The logistics of the psychological flow of information will emerge as the next competitive advantage – We will need to develop teams that are expert at teamwork, non-defensive management interactions and team chemistry – We will need to invest in the right blend of organizational capital elements Outcomes • How much operating budget can this save relative to competitors? 2x? • How much opportunity can it create relative to competitors? 5x? 42
  43. 43. Summary Computation and transport will continue to dance with each other • 20 years from now, computing systems performance will matter! • Denser imaging will require high performance computing data reduction and faster networking • Network closets will be come HPC closets adjacent to equipment quickly throwing off vast amounts of data • Computation and transport will merge and differentiate in many ways to deal with performance issues – E.g., disk architectures today are benefitting from ‘in-flight’ computation, desktop virtualization is exploiting local CPUs and graphics processors (wherever they are) to save on network latency and compute resources The individual and organizational dysfunctions are enduring • The knitting together of human/technical systems is the stuff of competitive advantage • The role of the CIO is likely to be strengthened and heightened as IT-induced competitiveness continues • The role is shifting to mastery of some abstract financial concepts and more difficult human dysfunctions, dynamics and teamwork skills 43
  44. 44. Final thoughtsAs information flows from a complex supply chain to a complexinternal value-network, between systems and human brains,and with many technical and psychological impediments andaccelerants, someone has to plan, monitor and nudge forwardthis strategic investmentWhy not the CIO? 44
  45. 45. “If you aren’t confused, you haven’t been paying attention!”– Tom PetersQUESTIONS? 45