We have several stages of dealing with Kali: denial, fear, death and rebirth. All of these industries have walked that path. The really odd thing is that, because IT is so highly technical, IT is still in the denial stage. We see the writing on the wall with the cloud and all the services it provides, yet compartmentalize it in our mind. It is only for a subset of things and won’t drastically affect me. We use justifications like performance and availability, data security, and even cost to protect the old ways. We convince ourselves that is both the truth and the logical choice. We deny the fact that Kali will touch everything.
The key to this internet change is instant gratification. I want what I want right now! Line of business leaders go to IT for business projects and see a wall of red-tape and process. They go to a service provider of either SaaS, IaaS, or PaaS and get immediate gratification. Are there ramifications to this behavior? Of course! I can see that clearly myself as I use Amazon and happily “One-Click” my credit card to its limit.
IT needs to move quickly into providing services for their business that provide instant gratification. IT will no longer spend the majority of their time managing the details of ‘things’ like network, hardware, storage and middleware. They will be managing services for the business. Many are starting on this path and it involves choosing tools that allow you to provide cloud (public, private, hybrid) management and automation to the business.
One way to provide added value over service providers is in the area of Application Performance, Availability and Event Management. Most service providers will simply provide contractual SLAs which guarantee up-time for their particular service, or perhaps some published metrics about the service. There is no information at all about how your particular custom use of that service is performing for you.
When choosing or evaluating whether to keep or replace existing tools for monitoring such apps, these are the most important questions:
- Does this monitoring solution support cloud-based applications?
- Can the solution itself exist in the cloud?
- Can the solution dynamically scale?
You would be surprised that many commercially available monitoring solutions, and especially legacy ones, will fail in one or more of these areas. The first two are obvious, but scalability is critical. As developers try to emulate the architectures of Web-scalable systems, the monitoring system must be able to dynamically scale to fit the app.
Web scalable systems are those where if we increase the resources in a system, it results in increased performance (i.e units of work) in a manner proportional to the resources that are added. Scalable systems need to be operationally efficient, meaning as you add more nodes to the system, it doesn’t require more operators because they are still performing the same tasks.
The key to the survival of IT is to begin competing with the service providers. Provide added value, yet keep instant gratification your number one requirement. Otherwise, IT may end up as one of Kali’s casualties instead of one of her rebirths.