.
ECS can be delivered as an appliance within the customer’s data center. With the ECS Appliance, the storage system is delivered, serviced, and prepared by EMC. Additionally, with the ECS Appliance, customers will get a single integrated hardware and software solution offering the best configuration for object storage environments. As a second option, ECS software enables customers to bring their own commodity hardware and run object and HDFS data services on their own equipment. With this option, customers benefit from both the low cost of commodity hardware while not compromising enterprise storage capabilities such as geo-replication, data snapshots, and much more.
In Q2 2015, ECS 2.0 was released free and frictionless. This means that end-users can download this version of ECS with unlimited capacity and unlimited license expiration. This allows customers to go from application prototype to production very quickly. They can add full license support when they are ready. The free and frictionless version of ECS is containerized allowing it to be deployed on a variety of Linux distributions.
ECS can be deployed in either a single site or in multiple sites that can then be federated together.
The basic building block for ECS is a single ECS appliance, which is also often referred to as a rack.
In a single site, you may configure one or more Virtual Data Centers (VDCs). Each VDC defines which ECS racks are to be grouped together.
In a multi-site deployment you will have multiple VDCs. ECS enables you to configure replication either within a single VDC or across VDCs. This provides flexibility in solution design allowing for data segregation, protection against many types of failures, and access from anywhere.
The basic building block for ECS is a single ECS appliance, which is also often referred to as a rack.
In a single site, you may configure one or more Virtual Data Centers (VDCs). Each VDC defines which ECS racks are to be grouped together.
In a multi-site deployment you will have multiple VDCs. ECS enables you to configure replication either within a single VDC or across VDCs. This provides flexibility in solution design allowing for data segregation, protection against many types of failures, and access from anywhere.
This diagram shows an eight node ECS Appliance. Each ECS Appliance contains several components: 1 and 10 GB Ethernet switches, Server Chassis, and Disk Array Enclosures (DAEs). The ECS Appliance can be configured with one server chassis and four DAEs or two server chassis and eight DAEs. Each server chassis contains four server nodes. Each node connects to one DAE
Servers: The top row shows a front and rear view of the Intel server. The server is a 2U high chassis which exposes internal disk bays associated with each server blade. Two of the slots are populated with 300 GB disks that are mirror images for redundancy. These disks are used as local storage for each processor and contain boot and operational code for each blade.
Disks and Enclosures: In the middle row we present two pictures of the standard Voyager Disk Array Enclosure (DAE). The Voyager DAE can be configured with up to 60 disks; each DAE is connected to a single node. Currently ECS configuration supports fully or partially populated DAEs with either 15, 30, or 60 disks.
Networking Switches: The bottom row presents ECS network switches. Each ECS rack includes three Arista switches. Two of them are 24-port 10 GbE switches that ECS refers to as “hare” and “rabbit.” They are used to provide redundant connectivity to the customer network and also as connectivity between the nodes. There is a third Arista 48-port 1 GbE switch that ECS refers to as “turtle.” It provides management connectivity to each node locally and also remote monitoring and management (RMM).
Disks and Enclosures: In the middle row we present two pictures of the standard Voyager Disk Array Enclosure (DAE). The Voyager DAE can be configured with up to 60 disks; each DAE is connected to a single node. Currently ECS configuration supports fully or partially populated DAEs with either 15, 30, or 60 disks.
Networking Switches: The bottom row presents ECS network switches. Each ECS rack includes three Arista switches. Two of them are 24-port 10 GbE switches that ECS refers to as “hare” and “rabbit.” They are used to provide redundant connectivity to the customer network and also as connectivity between the nodes. There is a third Arista 48-port 1 GbE switch that ECS refers to as “turtle.” It provides management connectivity to each node locally and also remote monitoring and management (RMM).
The infrastructure layer refers to the operating system, networking services, and some diagnostic tools. This layer enables the nodes within the appliance to communicate with each other. Each node is provided an external IP via the public interface and public host name. The ECS Appliance is configured to use services available in your data center including external DNS, Active Directory, DHCP, SMTP, and NTP.
The fabric layer refers to the data fabric, which is responsible for clustering the nodes together, maintaining the cluster health, and facilitating software installation and upgrades as well as alerting.
The ECS provisioning services is responsible for configuring the object personalities on the disks within the nodes. When the object service is installed on the nodes, the ECS portal becomes available for appliance management.
After ECS configuration is complete, end users are able to interact with the underlying storage system via the storage services layers. They can access data within the ECS system using any of the configured protocols including S3, Atmos, Swift, HDFS, or CAS
The fabric layer refers to the data fabric, which is responsible for clustering the nodes together, maintaining the cluster health, and facilitating software installation and upgrades as well as alerting.
The ECS provisioning services is responsible for configuring the object personalities on the disks within the nodes. When the object service is installed on the nodes, the ECS portal becomes available for appliance management.
After ECS configuration is complete, end users are able to interact with the underlying storage system via the storage services layers. They can access data within the ECS system using any of the configured protocols including S3, Atmos, Swift, HDFS, or CAS
ECS packages the essential software components in docker containers. This simplifies deployment and provides flexibility in the targeted hardware commodity configurations.
The diagram depicts the key docker containers that are activated in a functional ECS including Lifecycle, Fabric, Zookeeper, Registry, and Object. Please take a moment to review the primary function of each container
The diagram depicts the key docker containers that are activated in a functional ECS including Lifecycle, Fabric, Zookeeper, Registry, and Object. Please take a moment to review the primary function of each container
asdfaffs
ReplyDelete