TechInsight Magazine

Understanding Intel's Premium RAID Features

Intel's Premium Feature Keys are easily added to RAID controllers already in the field; simply drop the key onto a header and configure through Intel's software

When you buy a piece of hardware, you expect the features listed on that product's specification sheet—no less, and generally, no more. We were naturally excited, then, to hear news of Intel's latest RAID controller cards, which can be upgraded, even after they've been deployed, to enable a trio of impressive capabilities.

The seven models that support these three new Premium Features are all 6 Gb/s SAS- and SATA-capable, making them attractive upsells right out of the gate. After all, when your customers pay for connectivity on a per-port basis, doubling throughput means accommodating the same number of drives using half as many connectors.

Boosting Performance

But increasing link speed isn't the only way to supercharge the storage subsystems you build. One of the three Premium Feature Keys, called SSD Cache with FastPath I/O, actually covers two distinct usage models with performance-enhancing functionality.

Mainstream Value: Intel’s RAID Controller RS2MB044 offers four internal and four external SAS/SATA 6 Gb/s ports with support for all three Premium Feature Keys, making it a solid first storage controller with upgrade flexibility.

In the first, SSD Cache learns what data your customer uses most often and copies it to an installed SSD, which serves as second-tier cache. In read-intensive applications like Web serving, file serving, SQL, and Exchange, performance can be accelerated by as much as 50 times. Should your customer's working data set exceed the capacity of an SSD, Intel's controllers allow you to attach as many as 32 SSDs in a CacheCade cache pool. For most SMBs this is the way to go, maximizing value through a combination of flash-based drives and conventional magnetic disks.

Businesses that demand the utmost in performance from their storage configuration will want to deploy an all-SSD setup employing FastPath I/O. By increasing the transactional performance of applications with small and random I/O patterns, FastPath can increase performance by up to 200%.

A Higher Degree Of Protection

Of course, while performance is certainly critical to any RAID card's allure, most customers are principally concerned with data protection. Rapid Recovery Snapshot, another Premium Feature, facilitates the capture of snapshots and tracking of source volume data changes, making it easy to recover from an application corruption or malicious data deletion.

This is different from the overnight backups you're already running on customer data. Let's say you want to protect a boot drive specifically, allowing you to minimize downtime in any circumstance. By isolating the boot volume to a Virtual Drive and enabling Auto-Snapshot, that volume can be recovered in minutes using the last known good point-in-time. Configuring Rapid Recovery Snapshot is simple through Intel's RAID Web Console. Just drop the upgrade key onto your controller card and go.

Synergistic Security

Self-encrypting hard drives are quickly becoming the preferred technology for mitigating the vulnerability of storage at rest—disks being decommissioned or moved from one location to another. But the effectiveness of self-encrypting drives is limited to the tools used to manage them, and those tools have been an inhibitor up until now.

Intel's Drive Encryption Management key enables the company's controllers with management functionality, locking drives as soon as they're physically removed from a server or workstation. When a drive is powered back on, the Auto Lock feature requires authentication to unlock the encryption key and read data. In the event of a repurpose, Instant Secure Erase deletes the existing key, generates a new one, and consequently locks down any previously-stored data.

Depending on your customer's storage subsystem, Intel's Premium Feature Keys will appeal in different environments. The real take-away here is that it's now possible to sell the latest storage controller technology, and then follow-up with inexpensive functionality upgrades, without needing to replace perfectly-modern hardware. That's what we call SMB flexibility. For more information, check out www.intel.com/go/RAID.

Now’s The Time To Tackle Windows 7

Did you know, that later this year, your customers’ ability to downgrade their Windows 7 licenses to Windows XP will expire? Small and medium businesses clinging to the older operating system for fear of application compatibility issues in Microsoft’s latest and greatest need to start thinking about a remediation plan for legacy software holding back an upgrade.

Transparency Is Key: Applications running in Windows XP Mode are virtualized, but you’re able to launch them the same way as any other app.

The good news is, that in reality, there is very little to stress out about. Microsoft made application compatibility one of its top priorities during Windows 7 development, and there are a number of different approaches that you can take as a reseller to ensure a transition goes smoothly.

One of the most inspiring adoption tales comes from Intel, which is in the process of migrating from Windows XP to Windows 7, citing the new operating system’s performance, security, manageability, and productivity advantages as reasons to upgrade. Before it flipped the switch on more than 80,000 employees, though, it ran a beta deployment with 3,000. During the course of the test, Intel determined that roughly 25% of its Windows XP-era apps would require remediation (meaning 75% installed and executed properly).

Giving your customer a solution for that small fraction of problematic titles is what facilitates a smooth and timely migration. For Intel, it was important to establish a controlled workflow, where software versions could be consolidated, tested, and remediated. But the process is simpler in an SMB environment; it really boils down to a couple of steps.

32nm Debuts

Why is it that the enthusiast market always seems to generate the most excitement? Sure, record-breaking benchmark numbers are great, and high-performance hardware is always tons of fun to tool around with. But at the end of the day, it moves in lower volume. You want to get your customers excited by showing off the latest and greatest, and then present the solution in their price range based on the same architectural principles. Until now, that wasn’t possible. Intel didn’t have an entry-level incarnation of its Nehalem design. The situation has changed, though.

The Family Is Complete. Intel CEO Paul Otellini introducing Core i5 and Core i3. A shift to 32nm manufacturing enables the highest-performance mainstream processors ever seen.

The venerable Core 2 family has had quite a run as the crown jewel in Intel’s desktop portfolio, consisting of dual and quad-core models, 65nm and 45nm manufacturing technologies, and various cache configurations, bus speed settings, and thermal envelopes. In fact, nearly four years have passed since the company first introduced us to the architecture that de-emphasized raw clock speed and instead shone a spotlight on efficiency—getting more done in less time with smarter execution, all the while using less power. Core 2-based CPUs continue to offer an attractive balance between performance and pricing. However, it’s hardly a secret that there’s a new micro-architecture on everyone’s radar.

First seen late in 2008, Nehalem incorporated additional optimizations for power and performance. Intel’s transition from Core to Nehalem went particularly smoothly for the company. And no wonder—it had already adopted 45nm manufacturing (the “tick” in Intel’s “tick-tock” technology cadence) as an enhancement to Core back in 2007, perfecting the process before applying it to Nehalem. Thus, the “tock”—actually implementing the new architecture—scaled well. In quick succession, we were shown high-performance desktop platforms, dual-processor server configurations, and workstation platforms with tons of second-gen PCI Express connectivity. Intel deployed an array of quad-core, Hyper-Threading-enabled CPUs, addressing your highest-margin opportunities in one fell swoop.

But Nehalem was conceptualized to be more than just a monolithic quad-core die servicing the server, workstation, and high-end desktop segments. In fact, Intel deliberately designed the architecture to be scalable, which is why, more than a year after its introduction, we’re now seeing dual- and hexa-core derivatives employing the same exciting technologies your customers have been working with since 2008.

LSI JBOD's: Not Just Another Box

LSI 620J JBOD: The LSI JBOD 620J is a 2U, 24 drive unit supporting 2.5” drives.

Storage is a key pain point for today’s businesses. Small, large or in between, it doesn’t matter. Businesses find themselves constantly stretched to their storage limit and they’re looking hard at how to solve this dilemma. LSI’s introduction of their new line up of JBODs goes a long way to help solve this problem that is both easy to implement and cost effective.

The JBODs come in two varieties. The LSI JBOD 620J is a 2U, 24 drive unit supporting 2.5” drives. The LSI JBOD 630J is a 2U, 12 drive unit supporting 3.5” drives. Both feature Enterprise class construction with design considerations for vibration isolation and thermal control offering a robust enclosure for a wide variety of applications.

The JBODs feature an Environmental Services Module (ESM) canister that offers LED support for unit identification, service required, and overheat within the enclosure. As part of the drive LED support the ESM will also show power, activity, fault, and drive removal, the fault indication being a key feature that lets you quickly discover which drive is bad so you don’t pull a good one by mistake. Included within each ESM canister is a 6G expander that allows multiple drives to share each SAS port to maximize connectivity and performance.

The JBODs feature both redundant power supplies and the option to install redundant ESM canisters. The redundant ESM canister is a key value add as they enable the ability to support failover configurations which is critical in high availability environments.

The LSI JBODs are designed to work with a wide variety of drive types including SAS, SATA, and SSD devices. With the included expanders they have built-in support for both 6Gb and 3Gb technology, so the type of drives that can be used in the each JBOD is limited only by form factor and the host interface being used (The compatibility for drive usage is dictated by the host system controller or adapter).

The LSI JBODs are easy to integrate into existing server environments, connecting to the system a couple of different ways. You can use a host bus adapter like the LSI 9200-16e or 9200-8e, giving you the ability to access the drives directly or you can use a SAS/SATA controller like the LSI MegaRAID 9280-8e which provides RAID functionality to use with the drives. Both types connect externally from the server to the JBOD using industry standard multi-lane SAS cables.

You can add additional JBOD units as needed to scale out capacity with the ability to cascade multiple units together using the LSI expander in the ESM canister. For example you can start with a single LSI JBOD 630J, populated with 2TB drives – that gives your customer 24TB of available storage. Later you may decide your customer needs additional storage, using the cascading function of the SAS expander you can then add an additional 630J to the first unit to scale out the capacity to a total of 48TB on the fly without the need to shut down the system. This ability to scale is limited only by the amount of devices supported by the host interface.

Using LSI’s JBOD technology as a key part of your customer’s storage backbone means scalability isn’t an issue. Adding more storage, larger drives, different drives, all of that is so much easier when you’re using LSI’s JBOD technology.