IT Manageability Challenges in an Unwired World

Submit New Article

June 1, 2009 12:00 AM PDT


by Chris S. Thomas, George Moakley, Regina Lynn Preciado, and Matt Gillespie

Introduction

Hardware is already well instrumented, but the industry needs a new generation of standards to allow us to manage all levels of the solution stack.

Now that all large companies have developed complex, multi-site computing infrastructures, there is a tremendous level of opportunity to improve efficiency through manageability and instrumentation. A typical e-Business transaction involves a complex range of systems between the client and the back-end system that hosts the data on which the transaction is based, as in the following example:





Such complex systems of systems provide many potential points of failure and inefficiency. Moreover, enterprise management requires many discrete business entities to work together smoothly.

This article is the second installment of a two-part series in which Intel® Developer Services solicits opinions from two veterans in their respective roles at Intel to weigh in on topics of note. Another article provides a similar treatment of the topic of wireless security.

Intel Chief Strategist Chris S. Thomas is considered one of Intel's visionaries, charting future directions for industry and computing. George Moakley is Director of Enterprise Architecture at Intel and is the lead architect for Intel’s e-Business Solutions Lab. Their different business roles shape and inform their discussions, which together form a whole that is more than the sum of its considerable parts.

Chris S. Thomas on the Inefficiency of Systems Management

The first part of this article presents the strategic and philosophical point of view held by Chris S. Thomas toward addressing the enterprise-manageability challenges presented by widespread deployment of wireless technology.

There were years when we were still building the basic structure for IT and IT was hailed as a major “growth area.” Now, IT has become an expected expenditure, and the plumbing phase is giving way to the manageability challenge. With the wiring in the walls, enterprise leaders are applying their IT investment to improving their companies’ core competencies.

We have said that every business is an e-business. Now, every business is an e-corporation. Early attempts to link our internal solutions to external customers and suppliers has taught us that we had better automate internally and have our innovations in place to opt imize our own efficiencies, in order to deliver value to the others in the ecosystem.

In his discussion, Intel director of enterprise architecture George Moakley writes, “there is a tremendous level of opportunity to improve efficiency through manageability and instrumentation.”

In our new ecosystem of mobility and occasionally connected computing (OCC), manageability is both a tremendous opportunity and an enormous challenge for IT in the next five years.

Systems Management Is No Longer Efficient

Manageability is a continuum. In 1992 and 1993, Intel drove instrumentation into our hardware and software products so we could manage them. The capabilities are there, ready to be automated and used.

However, if we tried to use the old model of managing everything centrally, the number of things to manage would weigh down the entire world. In today’s business environment, we have multiple devices and multiple types of networks. We have (and Intel recommends) WMI interfaces to identify what networks are available and what performance is available on those networks before an application signs on. Applications use the existing instrumentation to self-manage their relationship with other applications and network capabilities.

In other words, we have moved the intelligence of management out of a central console and onto the solutions themselves.

We already can see this happening in the “updating environment” provided by vendors like Microsoft*, Adobe*, Macromedia*, and Symantec*, in which applications are auto-installing, auto-running, and auto-updating.

In the early 1990s, Intel's Andy Grove said that it took one system administrator for every 50 users. Now every one of us is going to become an administrator, because the applications we will be using will dynamically allocate resources as they need them.

Manageability in Flux: The Transition to Solutions Management

What we are advocating, and starting to see, is a transition from systems management to solutions management:

  • Systems management is looking at single devices, at uploading or downloading software, at individual hard drives, and at temperature in the server room – literally monitoring every individual element.
  • Solutions management is much more critical to the enterprise’s well-being than any individual box being up or down at any given point. For example, if email is down, nobody cares whether it is this server or that protocol that broke – they just want their email back.

 

This is especially true when you think about mobile computing and extended business-to-business connections where your customer, supplier, or employee is not always visible to you. They are part of the ecosystem, but you cannot manage them.

That is the biggest challenge we have to solve in order to realize true solutions management. As George says, IT managers “need a collective view of the messaging enterprise if [they are] going to management it effectively, and that effort is hampered today by the lack of consistent standards.”

How does one manage multiple boxes and solutions? How do we provide functionality, security, quality of service, and so on across the ecosystem, not just within an enterprise?

We have to leave behind a one-sided solution mentality and think instead, “What is my responsibility for the solution I am running? Where are the boundaries?”

End-to-end management is not owned by one company – there are numerous solutions participating in the equation. At the moment, we see little agreement on how problems are dealt with, who is responsible for each piece of the puzzle. That is the other challenge to surmount in this IT evolution.

Interoperability Is the Key to Solutions Management

Today, IT managers comprehend a solutions-management orientation, but the tools to make it happen are not there yet. Most IT departments are still wrestling with systems management, trying to deal with software updates on 400 servers and the like. In five years, IT professionals will be demanding solutions-management tools; right now, there are so few, in such early stages, that we do not have a large educated customer base.

We are starting to understand the need to take responsibility for how each IT shop contributes to the overall delivery of services in the environment. We are looking at a new manner of participatory management, the logical outcome of occasionally connected computing (OCC). We are establishing an asynchronous environment that breaks the paradigm of traditional management systems.

Every vendor will have to learn from this approach and start to be responsible for its impact on the overall ecosystem. As George says, “customers must demand interoperable solutions from suppliers. No one can sell me a comprehensive solution in terms of all the capabilities or all the product coverage I need.”

At that point, a reporting mentality will begin to evolve, and the industry will be able to create some standards.

What we are advocating and creating tools to support at Intel is a single code base that works across all architectures. There is common set of expected features and instrumentation across servers, clients, handhelds, desktops, laptops, and so on, so the ecosystem works more effectively.

This Intel model promotes volume economics. You do not have to rewrite code to deploy across multiple platforms because we provide this hypersegmented environment. Whether client, server, or massive server based on multiple Itanium® processors, the methodology and philosophy by which we manage solutions in those environments should not be different. That is one way Intel provides a lot of added value.

Listen Up: The Customer Knows Best

A solutions-management outlook creates opportunities for small- and medium-sized companies as well as the monoliths. The creativity of how we adapt to these changes will incorporate very large and very small companies – we will see a lot of innovation as we begin to comprehend what business flows look like in a wireless, multi-enterprise, OCC world. We will see end-to-end solutions that do not involve enterprises at all.

One key point to remember: Intel is not driving solutions-level standards as we have done with hardware and integration. We recognize that at the instrumentation level, it is our responsibility. But at the solutions level, the industry will need a common management philosophy in which we will participate and help to drive.

Each vendor has to participate as a piece of the ecosystem the customer puts into place. Customers expect this magical community of vendors to understand how they work with other vendors. At the solution level, customers know better what they need than vendors do. We just need to listen.

George Moakley on the Need for Higher-Level Manageability

The remainder of this article responds to Thomas's comments from George Moakley's perspective, that of a hands-on architect working on individual solutions.

Issues associated with distributed systems may be of a platform, application, or network nature. Various business groups (or even different companies) may manage devices and services, and it is likely that many systems do not communicate with one another at the management level. It is common for performance or availability issues in such systems to generate hundreds or thousands of events notifying support personnel whose systems are not actually involved.

Solutions that improve manageability of such complex environments yield return almost immediately, and they will continue to represent opportunity for some time going forward. Certainly, we will continue to see growth in this space for the foreseeable future, and most industries will need to invest in tools to improve systems-management efficiency.

On the other hand, the fact that manageability is becoming a mainstream, key concept in systems design raises the bar for this kind of product. When a new product or service comes to market, it has to represent real, measurable value to customers. Obviously, this development is good for the industry in general, since it tends to improve the quality of products and solutions.

Implementing Standards is the Key to Improving Manageability

Intel® architecture has done an excellent job of instrumenting platforms to make them manageable, which adds to the overall advantages in terms of total cost of ownership. From the customer's perspective, however, an industry-wide problem still exists in terms of managing large collections of machines.

In other words, while servers are highly manageable today, there is still a need for tools to improve management of the entire solution stack, including the software on those servers. The industry needs better automation tools to deal with all management entities on networks, whether they are applications, middleware, or hardware.

In the case of an e-mail infrastructure, for instance, I need to be able to look at those entities collectively as I determine why the mail is not going through on a particular day. As Chris has said, from the overall systems-management perspective, it is really a secondary issue whether individual servers are up or down, and I am not so much interested in that information as I monitor and troubleshoot the entire messaging enterprise.

I need a collective view of the messaging enterprise if I am going to manage it effectively, and that effort is hampered today by the lack of consistent standards. The entities on the network generally do not communicate well enough at the management level to provide this type of holistic view. I also need to be able to buy and deploy a solution without having to budget implementation costs that are equal to or greater than the cost of the tools themselves.

Finally, this holistic view also needs to support views of the enterprise at many different levels. That is, one needs to be able to group management entities in various ways, such as the equipment in a specific room or the applications associated with a particular solution, even if those groupings overlap with others. Tools to support that kind of flexible, collective view are missing in the industry today.

Many software companies have started making good progress in this area. Microsoft*, IBM*, and HP* come to mind, and there are others, but none has yet managed to create what I would call a complete solution.

Much of the problem has to do with the fact that standards have not yet been generally adopted among ISVs. The British ITIL (IT Infratructure Library)* standard, with which many ISVs including Microsoft and HP are providing good integration, holds great promise in this arena, although it is still too soon to tell what the future holds.

ITIL provides a great plumbing view for organizing environment support for services, and it could provide a robust basis for the next generation of solutions manageability.

Move from Managing Individual Systems to Managing Entire Solutions

In the late 1990s, day-to-day consumer focus in buying and deploying solutions was concerned with how to get the solutions installed and running. We are now shifting that focus to managing those solutions and to controlling the costs associated with them.

It is useful to distinguish between manageability and automation. Manageability makes management information available from a system through instrumentation. Automation puts systems in place to collect that information, synthesizing it into a meaningful form. One cannot effectively automate if equipment is not properly instrumented, and there is no point in instrumenting if no automation is in place.

Intel's role in this space is not to provide automation systems. Rather, we focus on making sure that our equipment is the best-instrumented available. We as customers must demand interoperable solutions from suppliers. No one can sell me a comprehensive solution in terms of all the capabilities or all the product coverage I need.

Interoperability must not be limited to batch uploads between systems. Tools have to talk to each other in real time, detecting events and generating automated responses. Those responses must be able to reroute traffic, start and stop equipment and services, and take whatever other corrective actions are necessary to keep the solution up and running.

Proliferation of Technologies Creates Management & Security Challenges

Proliferation of new platforms and technologies always creates special management and security challenges. For instance, the diversification of mobile applications and technologies provides tremendous productivity advantages, but widespread adoption demands that information can be transmitted securely and that the manageability of these systems is robust enough to control deployment costs.

As users move through multiple wireless modalities in the enterprise, such as from from 3G or 4G cellular networks to the WiFi network on a corporate campus, devices need to keep on working, and security needs to be seamless and transparent. Moreover, as a device rejoins the corporate network, we need to be able to ascertain whether it is still trustworthy.

A wide variety of instrumentation is required for these devices and their applications that is unique to mobile world. As an industry, we are just starting to figure out these challenges, as well as the opportunities for our products to distinguish themselves in this space.

Conclusion

The next generation of manageable solutions demands development of and adherence to standards by the industry at large. Wit hout consensus, we cannot succeed.

As a manifestation of how far the industry is from achieving this consensus, the next time you talk about manageability with someone, ask him or her to define it. In all likelihood, you will not agree on what the word means. That is how much of a disconnect exists in the industry about manageability and automation. From the customer's point of view, this issue needs to be resolved. That is the challenge and the opportunity.

Additional Resources

Intel, the world's largest chipmaker, also provides an array of value-added products and information to software developers: