Self-Healing Storage – Fail-Safe Storage and Heal-in-Time

This article was written by MSys Technologies LLC.

Self-Healing Storage – Fail-Safe Storage and Heal-in-Time

Consider the below scenario
You are a cutting-edge, cloud-powered ISVs. You have an esteemed customer base who are delighted by your uninterrupted services. On one fine day, you experience an IT outage. Wouldn’t your storage infrastructure suffer a backlash? OK. Like any other IT organization, you have a proper backup plan. This will ensure BC. But, just think again, what if this backup plan succumbs to failure? Now, you require adequate time to bounce back. Meanwhile, your customers get frustrated due to service interruption, which brings down efforts for a rich customer experience management.

We are almost entering the third decade of the 21st century, yet the modus operandi of technology is far from being proactive. Instead of waiting for something to break and then repair, we must unleash our technology prowess to undo glitches in the first place. So what is an ideal world for IT system administrators?

Imagine a moment of panacea, where your storage systems keep themselves healthy through self-monitoring. You would solace when they self-remediate their issues or possibly alert you in case of a complex surgical strike. Image, zero human intervention, while all these repairs initiated behind the scenes. This is a moment of panacea for IT System Administrators.

Roots of Self-Healing Mechanism
Autonomous Restore and Backup Repair
The consistent state of the storage system is to maintain data availability. But we know that metadata inconsistency can result in a nosedive. Therefore, we can induce self-healing mechanism via stimulation of the restore system. This will enable a device to capture the latest snapshot stored in a system, even if there is a restore failure. And this restore simulation procedure must be triggered periodically to ensure auto-pilot storage repair mode.

Post analyzing above-mentioned fallible storage possibilities, let us look at self-healing storage for maximum business availability, interrupted business operations and ultimately, delightful customer experiences.

Two Types of Self-healing Storage

1. Fail-Safe Storage
Consider you are building a pile of boxes. To achieve your goal, you must have a minimum of 15 boxes. Unfortunately, the ninth box is defective. In such a case, you can add one additional box to maintain the threshold and still achieve your desired goal. This is what we call a fail-safe method.

In storage engineering, this concept utilizes spare capacity to counter the challenge of hot swap or hot plug persistent in HDDs. You create disk drives, which has in-built additional drive capacity. Whenever there is a failure, it triggers an automatic swap to replace the failed drives.

A fail-safe storage architecture fits the bill for IoT applications. A typical IoT architecture will consist of a controller, storage server, sensor, and interface connectivity. The sensor collects the data to house in the storage server via interface network, which is relayed to controller resulting in a trigger action or communication. The hottest emerging IoT application is autonomous vehicles, which demands a high degree of safety. For the same, it relies on uncompromised data analytics. And this data is sourced from the underlying hardware integrated into the vehicle.

For example, Cypress Semiconductor Corp. launched’ Semper NOR flash family that is intelligently designed for fail-safe storage in ADAS. It has EnduraFlex architecture, which leverages ARM core for independently optimizing memory array. As a result, it can retain data for up to 25 years.

2. Heal-in-Time
Heal-in-Time means executing corrective measures in real-time. Upon detection of underlying issues in a software or hardware, it must immediately be contained. Consider Oracle Solaris Predictive Self-Healing (PSH). It’s OS consist of a Fault Manager daemon, fmd(1M) that continually runs in the background. If there is an issue, fmd(1M) will diagnose its nature by comparing with data of previous errors. It then assigns a Universal Unique Identifier (UUID) to it.

The Fault Manager daemon takes the software/hardware component offline, so it doesn’t affect the rest of the system. In meanwhile, it reports the affected components without requiring any manual intervention.

Similarly, The IBM XIV Storage System includes built-in mechanisms for self-healing to take care of individual component malfunctions and to automatically restore full data redundancy in the system within minutes.

Three Key Business Benefits of Self-Healing

1. Self-Maintained Systems
Machines require continual maintenance to be up and running. Similarly, a software application must undergo frequent up gradation to remain efficient. On occasions, caches must be cleared, and services must restart. This might sound a menial and easy task, but it eats out much precious time and hampers effective execution of their core competencies.

This is where self-healing storage can be fruitful. It would eliminate any kind of manual intervention and regularly update your application. It would ensure higher employee productivity along with seamless application functioning.

2. Intelligent Remediation
In case of any issue, self-healing storage would independently alert the IT admin department. Upon alerting them, it would intelligently remediate the issue. A self-healing store works like a tier-one repair machine – solving the problem of an application crash, disk overheating, excess disk vibrations, network connectivity, etc.

3. Cases beyond automatic remediation
Self-healing storage facilitates ongoing storage monitoring with real-time storage health insights. In a case, where the automatic remediation is beyond the scope of its capabilities, the issue can be immediately highlighted. Resultant – IT administrators get ready-to-act information about the problem with varied details. This ensures human action at the right time, which maximizes repair effectiveness and save.

Conclusion
Complex software development is the new norm. Datacenters will have to become equally sophisticated to support new-age application needs.

The rise of AFAs, 3D NAND, and hybrid infrastructure has likewise increased the usage of AI. This is because algorithms can be easily integrated into the arrays and other components of modem storage architecture.

The induction of self-healing capabilities within your storage devices empowers you to create a first-tier repair mechanic – automated, self-sufficient. A self-healing storage sucks up your technical overhead while generating significant cost savings. It monitors your devices/backup systems for any possible failures. Upon detection of any anomalies, it triggers a self-repairing process to avoid business hiccups and service outages.

Self-healing in storage sounds fancy for now, but let us be assured that soon it will be a necessity.