Instinctive storage simply performs its assigned task, no manual layout of LUNs or wizards to fill out. Similar to how your body responds to an increase in physical stress, you don't have to fill out a wizard to get your heart to beat faster or your lungs to bring in more oxygen. Instinctive storage responds automatically to the storage condition, or better yet to the application condition, at hand.



Legacy Storage


Traditional or legacy storage, that’s in place in the majority of data centers today, suffer from a number of issues. Among these are: a lengthy provisioning process, an underutilization of available capacity, a limited ability to fine-tune storage as the application demands change and an inability to effectively leverage the available storage tiers.


Provisioning Complexity and Process Issues

- Design and planning are time-consuming and require almost ‘superhuman’ knowledge of the physical array - and a spreadsheet full of layout data.

- The storage administrator becomes a ‘spreadsheet slave’ instead of a storage resource master.

- As an array grows, managing the performance of existing volumes becomes a substantial task.


Underutilization

- The act of provisioning is a (costly) ‘game of Tetris’, catching new volumes on free resources but inevitably leaving other resources unused.

- In primary storage, capacity and performance come at a premium. As a result, free space on primary storage has a premium penalty.


Inability to Fine-Tune Storage

- Cache settings are manual, static and often set at an array level.

- Workloads that may evolve from transactional to sequential create bottlenecks in performance.

- This "if it’s not broke" mentality inhibits optimization. The value of data changes and so should how it’s stored. But doing so, in the legacy model, is too costly and inherently risky.


Inability to Leverage Storage Tiers

- Moving data across from one tier to another requires data migration (planning, downtime, risk).

- This requires investments in specialized tools and storage platforms to accurately move and retain that information.

- In reality, because of scarce IT resources, it’s rarely a project that is actually undertaken.

Many storage managers believe that once soft costs are factored in, it’s more practical to buy more primary storage than to kick off an elaborate data classification project.



‘Automated’ Storage


To address the short comings of legacy storage, some storage system suppliers have added automation to their existing storage platforms. While sometimes these are new systems, often they are simply repackaged legacy hardware with the same complex software hidden by a wizard-driven interface. While these systems can improve interaction with the array, they still require attention from an informed administrator. This must be someone who’s done all of the design and planning, and is intimately familiar with the hardware and layout of existing volumes.


Given the lack of IT personnel and available time, there is still the likelihood of potential issues with identifying the need to use the wizard, identifying what specific information the wizard needs, monitoring to confirm that the changes the wizard made to the environment were effective and then preparing for the next time they run the wizard. Essentially a wizard only accelerates GUI navigation versus actually automating storage administration.


Identification Issues

- Identifying the issues in the environment is the single biggest challenge that storage managers face.

- While it is obvious when an application needs more storage capacity, determining where to get that capacity can be very time consuming and may lead to an unwanted purchase because the storage is locked into another application.

- Identifying performance issues is even more difficult and time consuming. Identifying that the storage has a performance problem, understanding what to do about that problem and how those changes will impact the rest of the environment can take months to understand.


Feeding the Wizard

- The act of feeding requires all the manual steps above to know what to put into the wizard. However, instead of going to multiple GUI screens or CLI commands manually, the wizard prompts you for the necessary information. Again, it speeds your way through the GUI, but does not provide intelligence on that storage.


Monitoring the Wizard’s Effectiveness

- After identifying and inputting all the changes to be made by the wizard, the actions the wizard takes must be monitored to see if those changes have a positive impact.

- After the results of what the wizard did are understood, the storage manager has to start the process all over again if the storage challenge was not resolved.



Instinctive Storage


The solution to the "storage is growing and the IT staff is not" problem is to look for storage systems that are instinctive, like utility storage from 3PAR. These types of systems respond to the needs of the environment and alter those responses as the environment changes. Instinctive storage then allows for a rapid provisioning process, achieves maximum storage utilization, can adapt storage performance to application demands and leverages available storage tiers. It does this all automatically, without the need for the user to identify issues, provide information to address the problem or monitor for performance improvement. In short, it just happens.


Self-Configuring

- After implementation the system configures, load-balances, and places data on physical disks automatically, without pre-planning, in order to easily optimize the performance and overall efficiency of the storage system. There is no need to go through the process of array design and LUN layout. Just start assigning capacity through application templates and the system tunes itself based on workload and available resources.

- Self-Configuration goes beyond initial setup and adjusts to the needs of the applications as they evolve. More storage is automatically allocated as required. Or if performance needs decline, the application’s data or parts of its data can instinctively be moved to a lower, and less expensive, tier of storage.


Self-Optimizing

- Performance needs of applications change over the course of time and even may have specific calendar driven periods where they need more or less performance. An instinctive storage system intelligently distributes and re-distributes data across different tiers or classes of storage based on user defined policies as needed in real time.

- The instinctive nature of these systems does not come at the expense of performance. While manual controls are available to reassure your DBAs, every performance test run with the 3PAR system as an example, have performed better than equivalent manually tuned systems.


Self-Monitoring

- Instinctive storage monitors itself – both physical as well as logical or virtual components -- so that IT professionals can focus on something else. While it does give reporting and the ability to interact with it to troubleshoot server and storage network issues within the data center, the automated system will respond instinctively to the health check data collected.


Self-Healing

- The self-monitoring aspect of an instinctive storage system also leads to a self-healing mode. Instead of having to bother the storage manager with the task, the self-healing system provides advanced error isolation and secure networking to centralized experts for diagnosis, escalation, repair, and guidance of on-site maintenance. In the case where a disk drive begins to show signs of problems, the system will notify the vendor support help desk and automatically move data to a new drive. Once complete, the instinctive storage system will alert the storage manager that the suspect drive is ready for replacement.

- This proactive nature saves the storage manager from dealing with performance loss and data risk while rebuilds are happening.


All of these attributes of instinctive storage greatly reduce the professional services and admin time required to get the system up and running. Bringing a system into production quickly and easily also improves confidence in that system which leads to quicker adoption.


Finally, an instinctive storage system also continually optimizes the investment. Most storage systems are rigidly set upfront, with little ability to adapt to a changing environment. An "if it’s not broke don't fix it” attitude is taken. The challenge is this often leads to work-stopping configuration changes and repairs. Reality is that the data center is a rapidly evolving environment and that continual fine tuning is needed. The only practical way to achieve this is through instinctive storage.

George Crump, Senior Analyst