Avoiding obsolescence

In IT, you’re either a manager or purveyor of obsolescence. It’s all about timing.

Following economic recessions, the winners are those who manage the loss of intellectual capital and resist self-perpetuating specialization, and learn to control, consolidate and converge technologies.

Are you a leader? Most are followers who proliferate pieces of technology while gambling on that piece getting lucky. Leaders will spot products out of step with the present business cycle by tracking rate of depreciation. They will reject increasingly ambiguous total-cost-of-ownership calculations and correct misalignment between business function and IT architecture.

Followers, on the other hand, become predictable profit streams for vendors, reliably upgrading at every planned obsolescence. Buying into convenience, the followers will accept free customization of their ERP and CRM systems, buy proprietary code and readily accept consulting/vendor partnerships.

Those who are neither clear leaders nor followers probably can spot proprietary technology lock-in ploys, but might overlook the total of all the distributed purchasing decisions: too much of everything.

My colleagues and I have learned to predict success and failure by comparing a company’s IT decision history against a benchmark chart of industry trends. We look for synchronicity between corporate technology path and industry’s ever-changing curves. From mapping our clients’ dead ends and obsolescence rates, we have derived five good practices:

1. ‘Moore’ doesn’t mean better.

Moore’s Law predicts that every couple of years a new generation of chip inevitably leads to new PCs, routers and miniature consumer goods offered roughly at the same price but with 100% more power. However, some leaders have rejected the necessity of following Moore’s Law with their purchasing. For example, Google CEO Eric Schmidt says the 64-bit Itanium processor doesn’t matter to his business. He doesn’t need maximum power at the higher density, but rather maximum functionality. Schmidt’s business is based upon algorithmic logic. Google’s transactional scale model relies upon lots of simple, inexpensive swappable boxes, and in the future, perhaps a distributed grid, not denser chips. Not all business processes benefit from the latest and greatest. Re-engineering processes to make sure they don’t require Moore maximizes ROI.

2. Resist the compulsion to outsource.

Outsourcing is like sweeping problems under the carpet. Most large companies outsource IT functions because they cannot resolve service needs efficiently. Rather than fixing the broken business processes and breaking down institutional barriers, outsourcing passes the challenge to the vendors. The lowest-priced outsourcer quickly magnifies inefficiencies with larger and larger invoices. It’s prudent to first reduce the number of vendors in an environment, create structures that audit business cases, standardize technology and offer incentives for adoption of cost savings through chargeback.

3. Consolidate with converging technology.

Optimization often paves the way for higher-level operations-support-system tools that integrate even the most proprietary approaches. For example, subdisciplines of IP traffic engineering, performance management, network monitoring, load balancing, perimeter security and chargeback are converging into a single optimized IP traffic management arena. These products share management information base or SNMP discovery algorithms. Alert thresholds and statistical methods can autoprovision and assist in end-to-end packet analysis.

Managers now can roll their own solutions with open source software, re-use their old probes or buy a single do-it-all box on-the-cheap. Finally, the rack and shelves of redundant tools can be cleared. With any luck, the department responsible for monitoring IP service-level agreements wasn’t outsourced to the IP vendor, because consolidation and control in the network operations center or data center with IP tool convergence can help save money through rightsizing telecom expenses.

4. Cut interaction costs.

Rapid exchange of information leads to greater productivity, but technology can create an information-exchange bottleneck. Some technologies require re-trained staff, and IT professionals might perpetuate obsolete systems.

The more staff dedicated to slowing exchange of information via content filtering, security policies and self-serving bureaucratic procedures, the higher the cost for every interaction.

To prevent information lag, train some generalists with practical, hands-on experience in multiple departments or encourage leaders to cross-manage a number of different disciplines. Creative cross-trained individuals will predict coming trends, think outside the cubicle and clean up interdivisional problems while speeding interactions.

5. Look past the Windows.

If we have learned anything from business-continuity planning post-Y2K and Sept. 11, it’s that we can’t rely on physical assets and soon-to-be-obsolete operating systems. Yet, Microsoft is struggling to proliferate the desktop software/hardware/location model as IBM tries to lure customers back to the proprietary central mainframe. Both paths perpetuate flawed code, foster dependence and worse: Your data might end up in obsolete archives with no functional readers.

Interchangeable, massively distributed but extensible architectures are likely the next phase in IT evolutionary thinking. The new and best standards in that arena are open. The future of proprietary Windows on the desktop eventually must cede to Internet-based graphical user interfaces and operating systems. Mainframe windows also will cede to extensible and distributed architectures.

Adhere to a principle of proprietary Windows nonproliferation – otherwise, it will fixate your architects and distract you from adopting the next global information paradigm.

Senior IT managers know change is the normal condition, but might not yet realize that tracking rates of obsolescence, slowdown of information interaction and proliferation of proprietary technologies will help them fend off moves to outsource their domains. Such a strategic realization defines the next business cycle’s leaders or followers; those to be and those not to be obsolete.