A FlashSystem grid is a new way of managing FlashSystem storage. It’s a federated scalable storage platform to manage grid made up of multiple FlashSystems and Storage Virtualise Controllers. It provides a platform and set of features that allows users to drive utilization of storage infrastructure non disruptively providing true private cloud capabilities such as scalability, agility, flexibility and simplicity.
FlashSystem grid provides innovative and efficient solution to solve data storage management challenges guarding against failures in data centre, preventing loss of data due to single data centre failures domain, adapting and scaling storage to meet changing requirements, keeping hardware and software modern to deliver performance, capacity and security, eliminating data silos so data is available whenever you need it, AI-powered data placement recommendations to improve infrastructure utilization and Federated single pane of management to bring ease of management and drive down cost of administration.
FlashSystem grid:
FlashSystem grid is a scalable storage platform that consists of multiple IBM storage FlashSystem or Storage Virtualize Controllers. It provides a single pane of management for FlashSystem grid deployments within the customer's estate. It offers a unique graphical user interface for visualizing a FlashSystem grid along with its storage systems and storage partitions.
Scalability:
FlashSystem grid will enable users to scale their cluster beyond the limitations of four I/O groups by introducing a new approach of loosely coupled systems within the grid. FlashSystem grid scale number of objects per system when I/O groups are added with linear increases in performance, capacity, and deployment objects.
Separate Update and Failure Domains:
FlashSystem grid has one FlashSystem (consisting of only one IO group) as a building block of the grid. So, multiple IO Groups in the grid can have separate failure domain. FlashSystem grid architecture doesn’t have restriction of any specific hardware type to be in the mix so different products from FlashSystem family can be a part of same grid. Architecture doesn’t mandate that all the systems should be at any one specific level of software either. Consequently, different systems in the grid can be at the different software levels and obviates the need of all systems in cluster to be upgraded at the same time. Separate failure domain and update domain is the key feature of this architecture providing users the required flexibility as they scale their infrastructure.
FlashSystem grid Construction:
FlashSystem grid is constructed with multiple FlashSystems. There are two types of systems that can be distinguished by their role.
Coordinator system: One system in the grid is designated to manage the construction of the FlashSystem grid.
Member systems: All other systems added to the FlashSystem grid are known as member systems.
User can do the management of FlashSystem grid from any of the system i.e. member system or coordinator systems. However, coordinator system is the only system from where FlashSystem grid construction activity can be carried out.
As user add the members in the grid, the trust is exchanged between these systems so as to allow FlashSystem grid operations across any of the system from any other system creating a single system view of entire FlashSystem grid.
Single Pane of Glass Management:
FlashSystem grid design provides unified management of all connected systems as a single entity through single pane of glass management GUI(Graphical User Interface). This GUI can be accessed from any of the system which is part of the FlashSystem grid. User can do the management of grid by logging in to GUI of any of the system i.e. member system or coordinator systems. This single pane facilitates monitoring and managing of all the systems and partitions in the grid from the GUI.
The FlashSystem grid dashboard, as shown in the screenshot, displays the total number of systems within the grid, health status , number of storage partitions, physical capacity. It also includes FlashSystem grid details such as the coordinator system name, FQDN or IP address, and FlashSystem grid UUID.
Managing all systems from Single Pane:
Visualisation of all the systems are collectively shown under the option “Storage systems” of FlashSystem grid dashboard. It displays the status of each system in terms of code level, last contact time, connectivity health , hardware health , physical capacity, total number of storage partitions and configuration problem.
To keep up with the ever changing demand for capacity and performance; FlashSystem grid dashboard provides enhanced user experience to add systems to the grid on demand. The same dashboard allows users to remove systems from the grid.
Managing all partitions from Single Pane:
FlashSystem grid provides an ability to create, edit and delete partitions on any systems in the grid. The Storage partitions page of the FlashSystem grid dashboard presents the list of storage partitions in the FlashSystem grid along with the information about HA status, DR status and migration status. Also provides the information about replication topology, volume groups and last contact time. FlashSystem grid enables user to start or abort Storage Partition migration between any systems in the grid.
Users can now run their workloads and configure various use cases like PBR, PBHA and 3-site PBR without worrying about required complexity of setting up each of the systems manually – many of those complex configurations such as certificate exchange, system configurations, pool linking, network configuration is automatically carried out by the grid.
Summary:
FlashSystem grid enables to revolutionize data management by eliminating time spent on workload migration so user can focus on more strategic initiatives. It optimizes capacity and performance of storage systems to achieve more, with low total cost of ownership. FlashSystem grid reduces effort to move to new hardware and software that is the reason of storage is always at its fastest, most secure and most simple. User never have to take planned outage so data is always accessible and admin don’t have to work overtime to keep storage running.
Reference material: