IInfrastructure Rebuild
Your enterprise's future success is dependent on its IT infrastructure "stack," which ranges from user-visible applications at the top to processor and networking widgets at the bottom. Overall, your infrastructure is obsolescing at a rapid pace, primarily because of improvements in "brute force" computing and networking.

To stay competitive, there are two approaches: 1) continuous small-scale improvements, which are hard to do, because everything is connected to everything else, or 2) widely spaced major top-to-bottom, wall-to-wall rebuilds, which are hard to do, because everything is connected to everything else. Indeed, there is no easy way forward, and for those who hate both the steady drip-drip of change or hate massive initiatives, the coming years are not likely to be comfortable.

Merely drifting via ad hoc investments also is neither comfortable or safe. The consequence of not exploiting new options is to fall off the  pace from a business perspective. Business innovation, process rationalization and process control all suffer from under-investment and inattention.

Background - The Technology Cornucopia

At the bottom of the infrastructure stack, hardware keeps getting more powerful, to the point that changes in operating systems and other capabilities are needed to harness that power. Moore's Law is plotted linearly on the adjacent graphic, to illustrate the giant leaps of recent doublings.

Currently, sixty-four bit is displacing thirty-two
bit computing, while processors themselves are being
packaged into "multi-core,"  tightly coupled two or four
processor combinations or, alternatively, they are being
incorporated into loosely coupled massive arrays and grids.
Virtualization - a capability dating back to the mainframes
of the 1970s - has been re-energized, so that today
prospectively every server and client device can be
configured to support multiple "virtual" machines. These
virtual machines can run different and even incompatible
operating systems and other software.

Disk resources continued to exhibit 30%+ CAGR in capacity and, like processors, are being decoupled from specific processors and becoming Internet addressable resources. Storage also can be "virtualized" and there are many new options for data replication and disaster recovery.

Very significantly, voice and data are converging with each other and with video, and all three with "mobility." User devices are becoming smaller, more versatile and less expensive to buy (although not necessarily less expensive to own). Expanding use of "sensing" - ranging from cameras to RFID - create new sorts of data and new sorts of workload on networks and storage.

At the application level, "loose coupling"  is replacing monolithic application architectures. Also, applications are gaining business process management capabilities, business rules are being
abstracted from applications to run on "rules engines," while assorted  services such as instant messaging, RSS information syndication and Bloogs and Wikis take on enterprise missions. At the heart of loose coupling and "service oriented architectures" (SOA) is reliance on Internet wired and wireless communications.

For more on the environmental changes that are driving change, click

Mind the Gap

If one compares your enterprise's current architecture with today's best practice, there of course will be a gap. Nobody can be at a zero gap, but the gap should be kept small. Although this problem is often discussed at the marco level (e.g., with respect to the U.S. falling behind), macro is merely an aggregate of many micro choices.

What is of concern is the likelihood that for your enterprise the gap is widening and perhaps is creating hindrances more fundamental than mere falling behind in software or hardware versions.

Analyzing the gap is therefore not some abstract "this technology is better than that" analysis that one can get out of an external publication, but requires a tailored view regarding the gap between your inftastructure "as is" and what it could be. The to be view would of course focus on enterprise priorities - e.g., more agressive adoption of wireless capabilities or more sophisticated handheld or virtual machines may tip the balance in terms of your enterprise being more competitive or being able to implement more cost effective business or information management processes.

It has become popular and comforting to suggest that change can be kept incremental and that there are no "big bang" projects needed of the1990's Business Process Reengineering sort. Indeed, incremental change is good, if guided by an overall, frequently updated view of the target "to be." However, it is often the case that to implement some competitvely significant, highly visible business process improvement, it is necessary to undertake some fundamental work and, worse still, some fundamental rethinking.

For further comments on the directions that exploitation of today's brute force computing might take with respect to information management, click

Management's Strategic Choices

The questions that the above technology trend presents to senior management are when and how to put this technology to practical use and what prerequisites must be met.

"How" is a difficult topic,primarily because of the connections and dependencies involved.  Coherence is important because the experts at a given level of the infrastructure stack - e.g., those most concerned with wide area networking or with server management or with mobile services - may optimize new capabilities in some way that injures another layer of the stack, or they may be pinned down by deadlocks that prevent improvement.

Also, the decisions made and the actions taken are not IT decisions, but "general" management ones that require a balancing of business process and priorities along with technical process and priorities. Today, IT is deeply ingrained into all business functions, so it is artificial and often counterproductive to address the matters within functional chimneys.

Strategic Impact on the Enterprise of Esoteric IT Choices

The day when business management made business decisions, leaving IT people to make supporting IT decisions is behind us. Increasingly, IT is the business, and general management must attend to key decisions.

It may be helpful to cite a particular choice to be made, which by its nature is strategic even though it involves a matter deeply buried in the IT "stack." That choice is when to migrate to IPv6.

There is the pending question as to when (some would say if) your enterprise is going to adopt IPv6 in some coherent, pervasive way, in place of today's IPv4. IP stands for "internet protocol" and version 4 is what its used almost universally today. IP version 6 is its somewhat overdue replacement. Given the subject matter, an IP version upgrade does not apprear to be a "general management" decision, but it certainly has general impac
t on all Internet-dependent enterprises.

Many readers are familiar with "area code" migrations that the U.S. telephone industry has had to implement to avoid running out of  telephone numbers. It has been necessary to create new telephone area codes,  not because of population growth, but because many more devices per person (and per employee) acquired modem capabilities - PCs, fax machines etc.

Somewhat similarly, among other advantages of IPv6 compared to IPv4, it greatly expands the familiar IP number range, eliminating the likelihood that any entity will run out of IP addresses. On the other hand, migration is a daunting task that has many transitional impacts.

IP addressing is strategic because devices ranging from vehicles to high-end RFID tags and plant production equipment are acquiring IP network connectivity. Additionally, there are other innovations such as the virtualization of  devices and software widgets that could expand demand for IP addresses faster than the increase in the number of physical devices. Indeed, the telephone number problem could soon be an IP address management problem as the entire telephony industry is gradually incorporated into the worldwide Internet addressing scheme.

Today, the boundary between your enterprises' internal internet (conventionally referred to as your intranet) and the external "real" internet is maintained by your firewall. At that boundary, a service named NAT (network address translation) translates your (often) single outside IP address into hundreds of internal IP addresses and vice versa. Enterprise internal IP addresses often fall into a  "non-networkable" range, while some may  not even be "street legal" because your enterprise reused someone else's IP numbering range.  Although NAT presently avoids conflict, NAT itself is in question, partly because the notion of an  inside and outside is itself challenged, while there are performance problems caused by address translation delays, especially on voice and real-time video.

It is something of a "no-brainer" that IPV^ must be adopted, but at the same time it will take a strategic, perhaps even Churchillian level of decision-making to translate that "should do" into actual accomplishment. To add to the difficulty, theIPv6  matter should be addressed in conjunction with other IT migrations. Therefore, implementing IPv6 would require some careful surgery, and it would probably be best positioned as part of a more general infrastructure "makeover."


The significance of this article to the reader is not whether to adopt IPv6 or virtualization or Voice Over IP or to replace SAP R/3 with SAP's new SOA-oriented architecture or to implement a BPM rules engine. All of those options and many more must be considered and, probably, adopted.

Instead, the article is intended to highlight questions such as whether 1)  you and your organization are aware that some profound decisions and commitments need to be made, 2)  your governance and planning capability can cope, and 3)  you can carve out manageable subsets of work in order to progress without risking the entire infrastructure. One challenge to every organization is to avoid "failure of imagination" - relying on a blind extrapolation of today's conditions rather than "visioning" an alternative future..

Organizations fall behind accumulate a change "deficit" and an ever-widening gap between what they are and what they should be. As the deficit grows,  there becomes a need to do a catch up effort. Therefore, most companies will have to do major, fairly compressed makeovers. In scientific study of evolution, the Punc Eq (punctuated equilibria) school of thought is that evolution moves in sudden fits and starts, and whether that was true for dinosaurs is open to debate, but it is certainly true for most enterprises.

Every year that goes by creates a greater gap and, hence, makes it more likely that a big push will be needed. Sooner is therefore better
To Site Map