“Having used DataCore’s SANsymphony solution across five years, we have established a stable, optimised infrastructure that responds to the challenges of our virtualisation journey. Our primary objective - to have full control of the storage layer to power the entire estate - has materialised and we now have the required flexibility, stability, performance and manageability to assure us for years to come.”
Working closely with DataCore Gold Partner, Park Place Technologies (formerly NCE Group), and with data sets of 250TB burgeoning 20% year over year, ICT went back to the drawing board to implement a detailed data management strategy facilitated through software-defined storage to provide performance, durability and high availability. Ross remembers when he arrived at Gruden:
“Then, the first priority in our IT estate was to consolidate and manage data sets by implementing a structured process and delivery plan. Had we not embarked on this plan, then the growth of unstructured data would have challenged both the scalability and stability of our entire virtualised infrastructure.”
With an environment of 48 physical servers and 16 physical vSphere ESX hosts running 22 virtual machines, Ross knew that his first process should be to create a stable, managed environment for data to reside within. Ross advocated this holistic approach in the knowledge that if the data storage was managed correctly, then Grundon would benefit from a more effective, scalable VMware vSphere® environment from which to migrate and provision new virtual machines (VMs). To achieve this, Ross upgraded his environment to the DataCore SANsymphony software solution, running on HPE servers.
More Control and Powerful Automation:
Park Place remained key advisors in the planning and implementation. Once installed, DataCore SANsymphony software quickly began to reap management rewards in real time, offering a granular view and management capabilities across Grundon’s entire storage estate of multiple Direct Attached Storage devices and arrays. DataCore’s Auto-Tiering function used the inbuilt intelligent heat-mapping feature to provide a configured view of system behaviour and performance allocation of data. Using this mapping, SANsymphony then automatically and systematically began the process of tiering data according to real-time usage. The ‘hottest’ data identified was automatically tiered towards Grundon’s fastest storage assets, (higher speed SAS arrays and towards Flash / SSD devices as they are introduced) with less-used data allocated to older and slower storage arrays, speeding performance for frequently accessed data sets and applications and utilising the most cost-effective storage tier. Better diagnosis of disk behaviours also empowers Grundon to spot issues before they occur -any latency issues are diagnosed and highlighted far quicker than through manual monitoring. Using the diagnosis, ICT can quickly determine whether any slow-down is attributable to disk issues or application issues. The software highlights areas where IOPS are bottlenecked to allow ICT resolution before users notice a decline in application performance.
Thin provisioning, or in Ross’ words, “absolutely allocating just the required exact amount of disk” formed the second part of the Grundon install. Decreasing waste is a key to Grundon’s core business. ICT were equally cognisant that up-front disk procurement based on assumptions about disk required had historically led to over-allocation. Deployment of thin provisioning immediately alleviated this overhead. Instead, thin provisioning created virtual volumes from Grundon’s shared storage pool to dynamically allocate more disk space – as and when required. Given this automatic availability of disk, I/O intensive applications running at intensive workload times - such as Microsoft Exchange Server, Microsoft SQL Server and Microsoft Dynamics NAV enterprise resource planning software (running in one Grundon division) - are supplied with I/O sourced from pools of unused disk, previously lying idle. If the system requires greater capacity, SANsymphony notifies the administrator that it’s time to add capacity.
Ross noted on the pooling and consolidation of storage resources, “Gaining control of our storage layer has had profound ramifications across the entire estate. It wasn’t until the SANsymphony platform was fully deployed that we appreciated its management fine-tuning capabilities.”
By using SANsymphony in a separated dual mirrored configuration with automated failover and failback at the Benson facility, the Grundon infrastructure continues to offer assured availability, with the software platform shielding applications from any failures at the storage device layer. This helps with planned maintenance and migrations. ICT simply work on one half of the mirror with the other half automatically taking over without disrupting applications. When the affected mirror is re-instated, the mirrored pairs are automatically resynchronized, the original paths are restored, and the normal dual node redundancy is automatically reapplied without ICT intervention. Full disaster recovery is offered at the Grundon Ewelme facility, some two miles away from Benson. Here, primary data is asynchronously replicated, so that in the event of a significant incident at Benson, Grundon have added assurance that the entire virtual estate can be resurrected within seconds.
Ross concludes: “With DataCore in place we now have the required flexibility, stability, performance and manageability to assure us for years to come.”
