ScaleIO is a software storage solution born in Israel and acquired in 2013 by EMC . The manufacturer launched version 2.0 of the software in the spring of 2016. ScaleIO is a block-based distributed storage technology that allows shared storage pools to be created by aggregating the disk capacity present on the servers contributing to the cluster.
Unlike most Software Defined Storage solutions, ScaleIO does not require all nodes to contribute to the storage pool to participate in the storage cluster. Some may just consume capacity. Another of the key strengths of the technology is that the nodes do not have to be homogeneous. They can use various OS (Windows, Linux, vSphere) and even offer very different storage characteristics, both in terms of capacity and performance (hard drives or SSD). ScaleIO offers integration points with OpenStack (Cinder) with VMware vCenter, with Mirantis Open Stack, but also with the volume management capabilities of Docker or Mesosphere.
Finally, if it is properly configured, the technology is able to load like no other. Several thousand nodes can thus participate in the same cluster according to EMC (although this is not necessarily reasonable). A 128-node 100% Flash cluster has recently been measured at 31 Million IOPS in 4K block reading and 23.7 million mixed IOPS - 70% read and 30% write.
ScaleIO 2.0: Enhanced Security and Resilience
The main novelties of version 2.0, according to EMC, focus on three main pillars: security, resilience and support for a larger number of environments. On the security side, ScaleIO now supports IPv6 as well as integration with Active Directory and LDAP. The different components of the solution also validate with each other to prevent an undeclared component from joining the cluster. Finally, the components now communicate with each other in encrypted form. Data at Rest Encryption is also available.
On the resilience side, the data is now subject to a checksum at the source and the target to verify that they have not been corrupted in transit before being written. It is also now possible to mark a node for maintenance, which triggers a redistribution of data on the cluster the time of restarting a node. ScaleIO then waits for the return of the node in the cluster to resynchronize the data which is much less consumer than a complete redistribution of the data. It is also possible to programmatically redistribute data to maintain full cluster redundancy if you want to retire a storage node.
ScaleIO now supports up to 5 nodes for managing cluster metadata. 3 nodes can be used for the storage of the metadata itself (including an active master and two slaves in stand-by mode) and two other nodes act as witnesses (or tie-breaker) in case of failure of the metadata nodes .
Finally, ScaleIO has been integrated with the Dell EMC Remote Management Infrastructure (the eSRS service for EMC Secure Remote Service). The technology can stream live cluster status to a Dell EMC control and monitoring center, which will assist customers in the event of an incident or alert them to a potential future problem.
ScaleIO 2.0 now interfaces natively with OpenStack, with CoreOS and Docker (via a special support mechanism) and should also do the same with VMware Photon OS . Ubuntu 14.04 LTS has also been added to the supported OS list (and a series of Juju charms are planned in the coming months to simplify integration with Ubuntu OpenStack). ScaleIO also has a RestFul API and integrates with Ansible and vCenter. A new web interface finally makes it possible to drive more simply a cluster (provisioning of volumes, management of performances ...).
ScaleIO 2.0 is available for free download on the Dell EMC website. The use of the software for testing or evaluation purposes is free of charge, and without limitation in terms of number of nodes or capacity. The acquisition of a license is required for production uses.
Unlike most Software Defined Storage solutions, ScaleIO does not require all nodes to contribute to the storage pool to participate in the storage cluster. Some may just consume capacity. Another of the key strengths of the technology is that the nodes do not have to be homogeneous. They can use various OS (Windows, Linux, vSphere) and even offer very different storage characteristics, both in terms of capacity and performance (hard drives or SSD). ScaleIO offers integration points with OpenStack (Cinder) with VMware vCenter, with Mirantis Open Stack, but also with the volume management capabilities of Docker or Mesosphere.
Finally, if it is properly configured, the technology is able to load like no other. Several thousand nodes can thus participate in the same cluster according to EMC (although this is not necessarily reasonable). A 128-node 100% Flash cluster has recently been measured at 31 Million IOPS in 4K block reading and 23.7 million mixed IOPS - 70% read and 30% write.
ScaleIO 2.0: Enhanced Security and Resilience
The main novelties of version 2.0, according to EMC, focus on three main pillars: security, resilience and support for a larger number of environments. On the security side, ScaleIO now supports IPv6 as well as integration with Active Directory and LDAP. The different components of the solution also validate with each other to prevent an undeclared component from joining the cluster. Finally, the components now communicate with each other in encrypted form. Data at Rest Encryption is also available.
On the resilience side, the data is now subject to a checksum at the source and the target to verify that they have not been corrupted in transit before being written. It is also now possible to mark a node for maintenance, which triggers a redistribution of data on the cluster the time of restarting a node. ScaleIO then waits for the return of the node in the cluster to resynchronize the data which is much less consumer than a complete redistribution of the data. It is also possible to programmatically redistribute data to maintain full cluster redundancy if you want to retire a storage node.
ScaleIO now supports up to 5 nodes for managing cluster metadata. 3 nodes can be used for the storage of the metadata itself (including an active master and two slaves in stand-by mode) and two other nodes act as witnesses (or tie-breaker) in case of failure of the metadata nodes .
Finally, ScaleIO has been integrated with the Dell EMC Remote Management Infrastructure (the eSRS service for EMC Secure Remote Service). The technology can stream live cluster status to a Dell EMC control and monitoring center, which will assist customers in the event of an incident or alert them to a potential future problem.
ScaleIO 2.0 now interfaces natively with OpenStack, with CoreOS and Docker (via a special support mechanism) and should also do the same with VMware Photon OS . Ubuntu 14.04 LTS has also been added to the supported OS list (and a series of Juju charms are planned in the coming months to simplify integration with Ubuntu OpenStack). ScaleIO also has a RestFul API and integrates with Ansible and vCenter. A new web interface finally makes it possible to drive more simply a cluster (provisioning of volumes, management of performances ...).
ScaleIO 2.0 is available for free download on the Dell EMC website. The use of the software for testing or evaluation purposes is free of charge, and without limitation in terms of number of nodes or capacity. The acquisition of a license is required for production uses.