The argument over agent-based versus agentless monitoring has been going on and been contentious for years. Initially, the power, reliability, functionality, and all-around robustness of agent-based monitoring overwhelmed the perceived advantages of lower cost, ease of implementation and maintenance touted in favor of agentless monitoring. Agent-based solutions had been around forever, or what passes for that in IT. They had a proven record of stability, utility and functionality that agentless monitorying just couldn't match.
However, all this is changing with the emergence of products from new and aggressive companies with agentless systems monitoring technology addressing these early weaknesses. At the same time, innovation and competition have driven agent-based solutions to resolve and alleviate identified problems. The result is increasing market interest in agentless monitoring and, more commonly, combined solutions using both technologies as enterprises look to save time and money in monitoring.
The traditional market for agentless has been among those willing to trade flexibility, breadth of reporting, robustness, multi-operating system support, and limited application availability for less expensive systems monitoring. Agentless systems monitoring was promoted for its ease-of-use, fast implementation and lowered maintenance costs and effort.
Traditionally, the reported strengths were never as clear and unequivocally present as advertised. Early agentless solutions were especially vulnerable to broken connections (preventing data collection), and the cost savings reduced when the lack of remedial and reporting functions necessitated purchasing additional tools to perform those functions.
Agentless solutions have come a long way from the earlier under-featured, less robust, bandwidth hungry initial implementations. It is not at all unusual to find agentless solutions that operate on all of the major operating systems, support a wide range of applications through open source and industry standard interfaces, have customizable tripwire settings, with report generation and analytics features.
Is agent-based technology consumption overstated?
At the same time, the reported weaknesses of the agent technology have proven to be far less pervasive and persistent than proponents of agentless technology asserted. Claims of excessive resource consumption are significant only with very low-end servers, centralized analysis reduces the local server load, and thin agents consume less memory.
The claims of demanding maintenance requirements turn out to be far less significant with automated, intelligent updating functions. Finally, agent-based tools isolated by a network loss can still perform many of their functions including data collection, local data reduction and analysis, as well remedial functions and report preparation.
Today, it is easier to examine the underlying technology to verify the latest advances in both agentless and agent-based technology. Side-by-side installation and operation of competing tools can also be used to determine the most efficient and effective combination of technologies to use. Examine features and functions of any candidate products to assure that technological problems as referenced earlier have been addressed. An internet search will yield articles and reports with data and information to help decision making.
Agent and agentless systems monitoring vendor offerings
One example of significant agentless solution is Longitude from Heroix. Longitude provides monitoring and reporting capabilities that both complements and competes with agent-based solutions. It delivers prepackaged monitoring for a growing number of application servers, web servers, databases, messaging applications, infrastructure and metrics for both end-users and business. Longitude provides the ability to create fully customizable reports, generate alerts, and displays that meet the specific needs of management and IT functions as well as the analytics to massage the data.
Competing with such aggressive new companies are the major players, BMC, CA, HP, IBM, Microsoft etc., all offer their own range of products that are making use of both agent-based and agentless technology. In a further variation, agentless technology is the backbone of Silverback Technologies, Inc. . Silverback offers their software to managed service providers to use as their technology engine to provide enhanced managed services to their customers.
A mix of agentless and agent-based technology does the trick
We are going to continue to sit firmly on the fence on this debate. Enterprises do not have to nor, in most cases, should they make an 'either-or' solution. Quite frequently, their individual needs will be quite well served using a mix of agentless and agent-based solutions. Every vendor of significant size uses a mix of agent-based and agentless technology. We see little likelihood of this changing any time in the immediate future.
Data center staff will have to understand their own operational needs to know which technology will best meet their needs and when. They will as well as take special care to perform due diligence to understand the implementation technology as they promote their solutions.
The trend in the market today is for the 'light-weight' agentless products to become more robust, powerful, intelligent and flexible over time. As mentioned, there are some highly attractive agentless tools available today but that is not true about all agentless products on the market. Let the buyer beware.
We believe that agents residing on and at the edge of the infrastructure will continue to be an attractive choice in many situations. This close to the action locations means that even when connectivity fails they continue to provide close to the source, comprehensive and intelligent problem detection and repair.
Continued automation, intelligent design, and competitive pressures mean that vendors must and will continue to lower the cost and increase the power of agent-based solutions. They will continue to be the preferred choice as long as network connectivity can fail.
ABOUT THE AUTHER: Richard Ptak is an analyst with Ptak, Noel & Associates. He has over 30 years experience in systems product management. Send comments and questions to: firstname.lastname@example.org This was first published in June 2007
This was first published in June 2007