[MUSIC] There's a really handy feature called a dynamic resource optimizer or audio which can continuously monitor the resource utilization of your systems in your environment, or in the background even when the system administrators are not working. Let's talk a little bit about this feature. So I just said that DRO can monitor your infrastructure in the background. What's the benefit? Well, when usage anomalies are detected, DRO can make migrations and changes to your infrastructure. There's another feature called capacity on demand or COD. This is a powerful tool that activates unused cores or mobile memory on a host, giving it more computing power or memory instantly as and when needed. On eligible systems, the power PC DRO works with COD to relieve overburdened systems. For systems that are not enterprise pull members, DRO migrate virtual machines. This animation here shows how DRO can activate mobile course as a host CPU utilization goes above a threshold you set. Initially, the host shown on the left is fine, but overtime let's say its CPU utilization goes above a value set. Then COD kicks in activates the extra cause and provides more processing power in the shared process of pull on that host. This is just one use case. Now let's see what happens when the enterprise pulls and COD option is not available. There are three hosts in this environment named hosts one, two and three. Host one has six virtual machines. Everything works well. Let's assume there's excessive demand for an application running on some of these machines and the hosts CPU or memory utilization reaches above a threshold value you have set. DRO can now enforce migrations if you have allowed it to. So, it identifies the best virtual machines to migrate to hosts two and three in order to get the utilization levels under control. Let's learn more about this by seeing how to implement it. As you may have guessed, this feature is available under the host group settings. Inside it there is a section for dynamic resource optimizer. You just check the box to enable the feature for a particular host group. Then you can specify if you want it enabled all the time or only during specific times. Also, there are two modes of operation. One is advise only, that is, you will only get messages regarding thresholds being hit, and recommendations DRO might have for you, but it won't actually implement any of the changes. I know several client administrators who use it in the advise only mode, or the other option is you can choose to operate it under the active mode. This mode will actively implement the migration changes it finds. Now let's assume that your business faces extreme loads during certain times, and you only want DRO to kick in during those times. Let's assume that the applications usually have a heavily used from, let's say, 3:00 PM on Fridays to 4:00 AM on Mondays. You can easily automate this. Just select the enable DRO during specific times option and set your desired time frame DRO will monitor your infrastructure during those times, and either advise you or actually implement the changes depending on your mode of operation. As you can see, enabling DRO is as simple as checking the box to enable DRO, selecting which mode, adviser active, you want DRO to run, and selecting which resource to monitor. We haven't talked about this yet. DRO can monitor CPU for all hosts, but can only monitor memory if the host in Noble is novalink managed. We haven't talked about novalink yet either, but just note that it's an alternative way to manage virtualization on your IBM power servers, but not hardware like the hardware management console does. So continuing on with enabling DRO, we have the need to select which optimizations DRO can perform, live migration, mobile core activation, or both. As we've seen, COD requires power enterprise pull support which can be purchased separately. And finally, let's not forget defining the runtime parameters of DRO. For example, CPU utilization threshold. There are a few more important settings here. First, we have the utilization threshold. This is the percentage level above which DRO should get triggered. Then the run interval. This is how frequently the host utilization is compared to the specified utilization threshold. Consider stabilization. This is the number of consecutive times at the specified utilization threshold must be exceeded before an optimization operation is performed. And finally, maximum concurrent migrations. This is the maximum number of migrations allowed in parallel within the host group. In this example, I have set up such that the 70% utilization threshold should be reached at least twice in a five minutes time frame for DRO to implement the migration action, and it is allowed to migrate ten virtual machines concurrently. Obviously, 70% may not be a great number to start migrations it depends on your use case. And finally you can see DRO in action in the hosts settings. Here we see an example sample utilization graph. The 70% threshold level seems to have been reached a few times, but they aren't within a five minute period, so no migration would have happened in this case. Now will continue on with our tour of power PC features with a look at improved volume lifecycle in the next video.
