Managing the Sun Ray business architecture

[Editor's note: Although Paul Murphy is offline until August 7, he filed a series of chapter summaries for a book in progress on Sun Rays and the Smart Display Architecture.]

There's an old saying to the effect that a fish rots from the head, but the reverse is also true: as a CIO you can't drive smart display adoption without first adopting key ideas yourself. This chapter is about that change.

That makes it the hardest chapter to write because I'll be dealing with intangiables and challenging the beliefs people have built their careers on.

• There's an introduction:

  The big thing here, the killer issue, the one big step most people never make, is the changeover from management to leadership. Think of management in terms of getting more hands on a job and then co-ordinating their activity: want a dozen workmen to lift a beam into place? that requires management: someone to assemble the workmen, tell them what to do, and then yell "heave" at just the right moment.

  Leadership is very different from that, it's about getting more brains focused on a job, and not at all about co-ordinating activity. If you're managing a group of workers they'll never get that beam in place without someone yelling "heave" - but a team leader doesn't have to be there to know that if a beam is needed, his people will get it in place.

  In fact a good leader is like a good sysadmin: someone everybody knows, but who seems to do nothing - except that things work after he's sat around doing nothing for awhile.

  Someone whose management style reflects a Wintel or data processing heritage can successfully implement a thin client solution for some or even a majority of desktops - but you cannot get to the smart display environment without changing the way IT works and therefore the way you work.

   

• There are the thin client organization benefits:

  "Going thin client" has significant organizational consequences for the CIO: you'll spend less time worrying about licensing conformance, auditability, or what users are sneaking into the office. More significantly, at least from a budget perspective, the people who repair or replace PCs can go away along with a good chunk of your help desk staff - and if you're big enough to have management infrastructures in both groups you can simplify some of those right out the door.

  The whole PC management, DHCP, and authorized load-at-boot business goes away.

  Identity management can be standardized and SOX auditor concerns can be more easily and naturally addressed -i.e. without pretending you know or can control what users do on, or with, their laptops.

  Downstream you can also look, again depending on how big your IT operation is, at simplifying networking - because PC to PC local communication needs are gone and the mantra that faster is better translates in networking terms to simpler is better.

  Basically these are all standard thin client benefits - you reduce your risk, your system complexity, and therefore your front line staff and can ultimately thin out the lower level management ranks accordingly. You do not, however, change the way you do business: rollouts go away, but software change processes don't actually change except at the trivial end of physical desktop testing - and neither does your reporting structure or your relationship with the corporate IT steering committee.

  Think of a thin client transition as a way of meeting your fundamental services provision mandate at lower total cost and you've got it: the desktop changes, a few people go away, but services don't change and neither do you.

   

• and then there's the change

  Things change dramatically, however, when user response to the IT controls that go with thin client and traditional thinking push you to adopt the smart display architecture - i.e. to counterbalance processing centralization with control decentralization.

  Right now, if you're a typical CIO your primary stance is that of the resource custodian or manager: discharging a corporate responsibility to provide a service at a contained cost. With a full smart display system that changes as you become a partner on the revenue generation side and cede directional control on the services side to user management. Basically this turns the traditional IT department inside out: instead of facing inwards as guardians of data and processing resources, you face outwards and push cycles and support at anything with a glimmer of user management support.

  Thus the nature of the CIO role changes: from custodian to evangelist; from management to leadership. When that happens your organization changes too: most of the people you have to manage now simply go away - no help desk, no PC networking, no PC repair or upgrade hassles; your middle management structure changes from one based on blocks of people filling narrow roles like "help desk manager" to one based on individuals working directly with other individuals. That changes your job too: you exchange "reports" for responsibilities as the people who substitute for you in direct employee supervision go away and your sysadmins morph into team leaders - meaning that your role becomes that of the facilitator smoothing their interactions with each other, with people selling you stuff, and with user management.

   

• Basically it's a whole different ball game and everything you already know is probably wrong:

   

  • with respect to server sizing and positioning (low utilization is good!)

     

  • with respect to recruitment, cross training, and team building (you want to eliminate job boundaries)

     

  • in terms of disaster prevention and organizational controls (e.g. service level agreements and disaster recovery plans make no sense, but you can generate substitutes on fly when these are needed to satisfy external auditors);

     

  • and in terms of appropriate software and run time models (application code becomes a throwaway good but everything comes down to the database);

  This is the "break it" chapter for most people - because getting users to control IT is counter to everything "we" know for sure - especially if we were raised in the data processing tradition or learned the IT management trade through struggles with Windows. It's the IT role that changes first, the methods change as a consequence of that - thus ideas like helping users fill niche functions with niche software are absurd in the traditional organization, but an everyday thing in the smart display world.

One major issue that probably should get introduced in this chapter involves the role of the external auditor as a drag on IT change. Auditors, particularly those with IT certification based on data processing standards, tend not to know much about IT delivery or technology and so try to impose processes that are decades out of date. Thin clients make it easier to meet these kinds of process expectations, but doing so sets up long term conflicts between IT and the user community and is incommensurate with smart display objectives.