Myriad Introduction and function

Myriad started working on a new project by ebay, MAPR and Mesosphere, and then forwarded the project to Mesos, "project development has moved to:https:// Github.com/mesos/myriad. " And then handed it over to Apache, it's a great project migration!

I. introduction of myriad (from concept understanding myriad)

The myriad name means countless or very large numbers.

The following is intercepted by the GitHub official website, translation level is limited which error also please advise.

1, Myriad is a Mesos the framework designed for scaling a YARN cluster on Mesos.

1. Myriad is a mesos framework designed to expand yarn clusters on mesos.

2. Myriad can expand or shrink the resources managed by a YARN cluster in response to events as per configured rules and PO Licies.

2. Myriad can expand or scale resources for each configured rule and policy to be managed by the yarn cluster.

I understand myriad and its role:

                

                650) this.width=650, "src=" Http://s1.51cto.com/wyfs02/M01/7B/08/wKiom1bFZKbTOCUUAAGyGmMGB5E654.png "title=" How-it-works2.png "alt=" Wkiom1bfzkbtocuuaagygmmgb5e654.png "/>

                

                myriad enables the integration of resource utilization and resource management across data centers when using Mesos. In this case, the workload of yarn is running on a shared cluster, which is more dynamic and resilient than a separate yarn cluster. This approach also enables data center maintenance teams to expand their resources to supply yarn (or take it away from yarn) without having to reconfigure the cluster.

        The origin of myriad

               Compare fire two Unified resource management and scheduling platform yarn and Mesos, comparing their two words there is a lot of it is not much different. I'll just divide the next use scenario:

                yran: When a job request is submitted to Yarn's resource manager, yarn evaluates the available resources and places the job in the appropriate location. It's a matter of deciding where the homework should go. Yarn is not designed for long-term service, or for short-life interactive queries, although it is possible for yarn to dispatch other workloads, but this is obviously not an ideal model. when there are no big data tasks in the queue, these resources are often not fully used. When a big data task runs, these resources are quickly used to the limit and are requesting more resources.      

                mesos: Using the two-level scheduling mechanism, where resource requests and offers are for frameworks rather than jobs, the framework can be seen as an application running on Mesos. Mesos's master node decision provides a resource for each framework, and each framework then makes decisions about which resource requests it can accept and which applications can run on those resources. When nodes in a cluster share multiple frames, this resource allocation method can obtain the approximate best data locality.

                 When you evaluate how to manage a data center as a whole, use Mesos to manage all the resources in your data center while using yarn to manage Hadoop tasks securely, but it does not have the ability to manage the entire data center. Data center operators tend to divide clusters into different regions (Hadoop clusters and non-Hadoop clusters) to address these two scenarios.

using Mesos and yarn in the same datacenter, there is currently a need to create two static partitions in order to benefit from the resource manager. This means that Mesos cannot function when the specified resource is managed by the yarn of Hadoop. This may be too simplistic, though it does work. But in essence, we want to avoid this situation.

can enterprises and data centers benefit from the coordination of yarn and Mesos? The answer is Myriad.

Myriad Introduction and function