Application Management and Application Aware Networking – Part 2
In Part 1 of this blog, I described the information gathered from Application Aware Network Performance Management (ANPM) solutions, as a “key enabler for the increasingly automated management systems of the future”. This blog entry explains why I believe this to be true.
Virtually every company with on-premise IT (as opposed to outsourced IT) has systems management tools in place. These are specialized toolsets utilized by support teams to monitor availability and performance of core hardware and software systems such as servers, databases, or virtual machines. The problem is that too many companies stop there.
For nearly 50% of companies, such tools are also used to (try to) monitor applications. This gives rise to the classic scenario described in virtually every application management webinar. A user calls about a performance problem, while the windows servers, databases, and applications servers all show “green” in the monitoring tools. In other words, the infrastructure is “healthy”, but the application still isn’t working.
Why? Because application execution is not simply a matter of what happens on server-based technology. It is equally a matter of what is happening in the connections between the servers– in the network. The network is the “lifeblood” of the application. It is the communication system that enables messaging and data to flow between tiers. For today’s complex, multi-tiered composite applications, these network flows can be incredibly complex. As a result, few companies have a precise understanding of how applications interrelate, who is talking to whom, and the path any one transaction takes as it traverses the data center.
This lack of visibility makes the troubleshooting process so difficult that application-related problems are often solved by workarounds– or just chalked up as “glitches” and ignored. As a result, the number 1 application management problem reported by IT organizations is “intermittent problems of unknown origin.” Problems aren’t solved, and they keep repeating.
ANPM solutions view the application execution from a unique vantage point. Strategically positioned across the execution environment, they “watch” network interactions and messaging across applications, transactions, tiers, data requests, and user requests. By exposing the complex communications within application ecosystems, they provide a foundation for a host of management and governance capabilities that could revolutionize application management.
Packet analysis and correlation could reveal the inner workings of even the most complex application environments. Packet headers, for example, contain “to” and “from” information. Massaged by the right analytics, such information could automate the process of “mapping” transaction flows from origin to endpoint– and of exposing interdependencies between hardware and software components.
Packet payloads contain data and messaging information. Analysis can reveal which systems are generating, changing, and ultimately consuming data. Eventually, this could provide a real-time alternative to traditional, warehouse-based Business Intelligence solutions. A few companies are already doing real-time analysis of flow information to track the value of the day’s sales transactions in real time, for example.
ANPM can be seen as one more unique source of information relevant to the overall task of application management. Consolidating, correlating, analyzing, and reporting across a variety of sources is the task of overlay analytics products. There are a variety of vendors incorporating flow-based information into real-time models of execution environments, which in turn support troubleshooting, root cause analysis, and change control. Vendors profiled in EMA’s recent Application Discovery and Dependency Mapping report, for example, include IBM, HP, OpTier, ASG, VMware, Service-now.com, and ManageEngine. The list of vendors included in EMA’s ANPM report is too long to cite here, but is available at: http://www.enterprisemanagement.com/research/asset.php?id=1801.
Moving from the primarily manual application management methodologies of today to the increasingly automated methodologies of the future requires a unifying force. This is where the promise of ANPM becomes most compelling. The information gathered by ANPM products may be the missing link that can make this a reality.