ROBO Solutions – The SDDC Perspective: Part 2 VMware vSAN from a ROBO context
What makes VMware vSAN the ideal technology candidate for any ROBO Solution?
Remote office environments often require shared storage whilst also trying to lower the total cost of ownership of the environment. vSAN simplifies performance and availability to storage policy attributes that can be applied to a single VM or groups/all VMS in a cluster. This remote vSAN ROBO cluster can be managed centrally and does not require a local admin resource and can be managed by your vAdmin which avoids the need for a separate Storage Administrator to manage the system.
vSAN delivers a shared pool of storage by aggregating local server disk storage in to vSAN disk groups on each host. The disk groups are then pooled together to provide a single Datastore of shared storage. No separate Storage Array System and dedicated SAN Infrastructure is required. As a result, from a ROBO perspective a vSAN solution can lower the TCO considerably when there are numerous sites involved as HCI Solutions can share existing local network switch resources (but be logically segregated) and requires investment in just the additional server hardware.
The data storage protection applied to VMs is based on a vSAN Storage Policy Based Management (SPBM)Policy. vSAN SPBM simplifies the operations for the vAdmin as storage configuration and management detail is abstracted so that the VMs storage availability and performance attributes can be applied intuitively with just a few clicks of the mouse. The vAdmin can also standardise ROBO solutions to have consistency across the sites by applying the same vSAN policy to all sites ensuring each vSAN ROBO Cluster has the same policy attributes. In its simplest form the protection applied normally allows a failure of at least one host (FTT=1), and it requires three hosts to implement this with a primary mirror object, secondary mirror object and witness (to arbitrate) object on a separate host. The additional of a fourth host enhances the solution as it enables the rebuild of impacted objects in the event of a host failure. The vSAN Logical illustration below puts this in perspective:
Thereafter the required size and scale of the cluster is dependent on the size and performance requirements of the workload being managed. The solution is then designed to meet those requirements. vSAN HCI Solutions are capable of scaling out as well as up, and this ensures flexibility in being able to deliver differing amounts of compute vs storage capacity by altering the node components to match the design requirements. There are many additional factors which are taken into account in a HCI Solution Design, such as: Business Continuity Replication versus vSAN Stretch Cluster, advanced features such as De-Duplication and Compression, Erasure Coding (Raid-5 and RAID-6 capacity optimised protection) as well as vSAN Data at Rest Encryption.
What vSAN architecture solution will bring all the vSAN simplicity, efficiency and policy-based management goodness, with the smallest hardware footprint?
With Remote Back Office locations in mind, where reduced workload footprint and lower TCO are expected, there also exists the option of a 2-Node vSAN Architecture. FTT=1 (Mirrored) is possible by dedicating witness objects to an external hosted Witness Appliance, with the primary and secondary mirror objects hosted on each of the hosts in the 2-Node vSAN Solution. The Core DC vSphere Cluster can be leveraged to house the vSAN Witness Appliances (VMs) and the Core DC vCenter can manage the Core DC vSphere Cluster and the vSAN ROBO site Clusters.
A 2-Node vSAN Solution can simply be 2 vSAN Ready Servers (VMware HCL approved vendor fixed hardware server specs) that the Solution Provider such as Asystec orders, plans, designs and deploys. This can be viewed as a simple procurement mechanism but what Asystec has found is that an Integrated HCI Appliance solution such as VxRail, jointly developed by DellEMC and VMware, provides a more robust and better ROI for the life time of the solution. VxRail is a fully integrated, pre-configured and tested hyper-converged infrastructure appliance, powered by VMware vSAN technology that simplifies and extends VMware environments. VxRail provides full hardware and software lifecycle management with a single click upgrade. Unlike a vSAN Ready Solution, VxRail provides and maintains the full hardware firmware and VMware software stack as a single build package that is fully tested as an integrated stack before release.
As of VxRail 4.7.100 release, a 2-Node Solution Architecture is available and can address the ROBO business objectives of simple, consistent deployment policy as well as meeting the agile requirement with quick automated deployments. Below you can see a sample Solution Overview that pulls this vSAN and VxRail ROBO solution together.
Coupled with the appropriate level of vSphere and vSAN licensing (as discussed in part 1), the solution can provide a commercially compelling, as well as a technically compelling, ROBO solution.
Standardising storage policies (in software) across ROBO sites and solutions can be achieved very effectively with vSAN ROBO Solutions whether this be with a Two Node Architecture or with a larger vSAN Cluster node count.
Another important area that customers wish to standardise and manage easily is networks and network security. The ability to apply these policies in software is quickly becoming commonplace among enterprise strategy. This is equally, if not more important, at the edge in ROBO Solutions. VMware’s Software Defined Datacentre solution in this network and security context is VMware NSX aka Network Virtualization. Make sure to read part 3 in this 3-Part ROBO series for more details!