We are at an unprecedented place in the history of the computer industry. Both technical and economic forces are at work to drive innovation at an unprecedented rate and scale. These forces are creating new types of uses of IT services, new applications, new platforms, new infrastructures, and new models for selling and buying management software. Any form of management software that cannot address these new environments, and keep up with rate of change will die. It is therefore inevitable that management frameworks will die.
The Technical Forces Killing Management Frameworks
The diagram below depicts that main technical forces that make management frameworks into legacy products. These forces can be summarized as consisting of rapidly arriving and changing applications (Agile Development and DevOps), diverse software platforms, abstracted resources (the Software Defined Data Center), and distributed execution environments (scaled out hardware, distributed data centers, hybrid clouds, and public clouds).
Technical Factors That Kill Management Frameworks (click to expand)
The specific technical issues that the above factors create for the viability of management frameworks are:
- Management frameworks are revealed by this rate of change to be a flawed concept. The concept behind a management framework was that one product from one vendor could manage everything that was involved in delivering your IT services and the business services supported by those IT services. It turns out that no single vendor can keep up with the rate of change, and they have not been able to for quite some time.
- Management frameworks are out of date, and they cannot be brought up to date. IBM, BMC, HP, and CA will not be able to evolve their frameworks to meet the new requirements outlined above, while simultaneously broadening their scope to deal with the new things that need to be managed and re-architecting their solutions to deal with distributed environments and clouds. HP at least deserves credit for recognizing this, and deciding to invest in building an entirely new set of management offerings targeted at these problems.
- While framework vendors all acquired Java APM products 10 years ago, those products were not built for a world of Agile Development or a proliferation in languages and platforms (PHP, Ruby, Python, Node-JS, Spring, CloudFoundry, and PaaS clouds). The APM components of the frameworks are therefore seriously out of date, which makes the frameworks out of date. Customers with modern languages, modern platforms, and modern deployment environments should go with APM solutions like those outlined in our post on SDDC Application Performance Management.
- The most rapidly growing market for infrastructure management – Operations Management for Data Center Virtualization, has gotten to the point where VMware has over 500,000 customers, and those customers by and large do not use frameworks to manage their virtualized environment. Rather they choose modern Operations Management solutions like those outlined in our post on SDDC Operations Management.
- Customers running applications across any combinations of private clouds, hybrid clouds and public clouds do not use management frameworks to manage either the infrastructure or the applications. This is because frameworks simply cannot be used to manage assets in such a distributed and abstracted set of environments.
The Economic Forces Killing Management Frameworks
The diagram below illustrates a fundamental change in the way management software is bought and sold. In the old way, an enterprise sales rep along with lots of technical pre-sales support gets the customer (the beneficiary of a variety of inducements along the way) to sign an ELA for an entire product family of management solutions. The teams that actually have to use all of the tools that have just been purchased are generally not consulting on whether all of those tools actually meet their needs. They then go around the ELA to buy the tools that actually do meet their needs.
In the new way of buying and selling management software, vendors make it easy for customers to try their products in production for free before the purchase conversation starts. In other words it is up to the vendor to prove that their products work before a sales conversation gets underway. Furthermore since the cost to the vendor of selling management software in this manner is so much less than the old way (there is no jet fuel in the sales cycle in the new way), prices of management software tend to be much lower in the new model.
Recommended Customer Actions Regarding Management Frameworks
Given that management frameworks are expensive and outdated legacy software products that do not meet current needs and that stand no chance of being enhanced to meet current and future needs, customers should take the following actions regarding their legacy management frameworks:
- Management frameworks should not be used to manage data center virtualization (VMware, Hyper-V, KVM, etc.), private clouds, hybrid clouds and applications deployed on public clouds.
- When servers get virtualized, legacy framework management agents should be uninstalled as a part of the virtualization process.
- New Operations Management solutions that are specifically built to deal with the unique aspects of data center virtualization, the software defined data center and clouds should be purchased.
- As more and more of the data center becomes software defined, the scope of what is managed by frameworks should be dramatically narrowed over time.
- Every time you renew a maintenance agreement for a framework, your goal should be to make the cost go down. You should not agree to renewals that last more than one year, due to the rate of change that the framework will be completely unable to keep up with.
- Once you implement your software defined data center you should to what Amazon, Google, Facebook, Netflix and other leading edge companies do – which is manage your environment without any legacy management frameworks at all.
Agile development, DevOps, language proliferation, platform proliferation, scaled out data centers, and applications distributed across private, hybrid and public clouds make management frameworks into dangerously useless legacy software. New economic models for selling and purchasing management software make frameworks unacceptably expensive. For these reasons frameworks need to be put on the glide path to oblivion.