[IMAGE]11936[/IMAGE]One aspect of the converged IP market that has seen the most rapid change is storage. Continuous advances in both the speed and size of storage media, as well as the reduction in cost, have created a moving target that is driving our industry toward more and more storage capability for systems of all sizes.
In this first installment of a three-part series covering storage types and related issues, we’ll take a look at some of the foundational principles of data and video storage. We’ll also touch on what’s around the corner in terms of future technology evolution and increased efficiencies.
The Bigger the Better
Of all the advances in digital storage, capacity has certainly moved the fastest. It wasn’t that long ago the size of a hard drive was measured in megabytes. Today we consider gigabytes as insufficient for all but the smallest installations, and terabytes are now the standard. Larger systems are already moving into the world of the petabyte (1,024 terabytes!). This acceleration in storage capacity was predicted by Moore’s Law more than 30 years ago.
Moore’s Law is a concept introduced in a paper by Intel co-founder Gordon Moore in 1965. It laid out the idea that computer hardware — originally it applied to the number of transistors on an integrated circuit (IC) chip, but has since expanded to other aspects of computer technology — doubles in capacity every two years. Moore’s prediction has been shown to be eerily accurate throughout the past couple of decades.
It is important to remember, however, that along with large capacity hard drives, the size of the video files themselves has also increased, especially with megapixel cameras. It is very tempting to simply think that “more is better” and devise a storage scheme accordingly. But it is still critical to do as much calculating as possible when it comes to determining storage requirements. Here is the basic forumla: captured image size + number of images per second + desired retention time = storage needs.
It’s enticing for some people to believe they can buy a recording box of some kind (DVR, NVR or hybrid) without hard drives and instead buy off-the-shelf storage devices from their local or online PC supplier. While the cost savings may seem to justify this approach, it is very important to keep in mind that not all drives are created equal.
That $88 2TB drive from the online PC store might work fine for your home computer or server, but that doesn’t mean it is appropriate for a video recording system application.
The PC and the NVR, while similar in hardware, have very different work patterns. Your PC, as busy as it might be, can’t compare to the duty cycle of an NVR or DVR. When using a home computer and its hard drive, you might call up documents or other information to read, you might write a document, you might check some E-mail or download some files off the Internet. All of which is pretty basic stuff; the percentage of actual time the drive is in use is relatively small. A few reads here, some writes there, but mostly downtime.
For the hard drives in a video recording device, the world is very different. Remember that in most cases a video system is working constantly, 24/7. Depending upon recording schedules and other factors, a drive can be in write mode 100 percent of the time. At the same time, with no interruptions, that same drive can be asked to read or retrieve video for someone’s review. That constant 24/7 writing with no breaks can take its toll on a drive.
In the past few years, hard drive manufacturers have seen the increase in the security market, and acted accordingly. One of the things they started to do was to test and assess hard drives to the more stringent video standards and engineer the devices to stand up to that level of punishment. Of course, these new standards eventually find their way into drives of all levels. Rest assured that the $88 2TB drive mentioned earlier doesn’t meet those expectations. In short, buyer beware.
Surmounting the Weakest Link
Hard drives are generally the weak point in any video recording system. The constant workout described above, coupled with the heat generated by that work, makes it very easy for hard drives to fail. Add the word “inexpensive” to that combination and you’re asking for trouble.
Keep in mind there are a couple of different ways the storage system on a DVR or NVR can be configured.
At the most basic level resides JBOD or “just a bunch of disks.” JBOD has no redundancy whatsoever. As video files are being captured they will be written to a single drive until it is filled up. The recording then moves to next hard drive and so on until all the drives are full — at which point the process starts again.
In the case of a JBOD box, if any single drive fails all the video is lost from that particular drive only. Understand that if this happens, the video is pretty much gone forever, unless the not-at-all-inexpensive services of a data recovery company are enlisted.
If there is a necessity to have the video around when you need it (and show me a video security system that doesn’t need that requirement), then there are several steps that can be attained to achieve that goal.
RAID or redundant array of independent disks is a system of utilizing multiple hard drives in a single machine to create a single virtual volume the recording box will use to store video. In essence, the operating system doesn’t know that the big storage volume it sees is actually made up of multiple physical hard drives.
RAID technology itself is broken up into different levels, each offering different benefits to the system. Following are brief explanations for these differing types of solutions:
RAID 0 — This, the most basic RAID level, offers the ability to create a large storage volume from multiple disks with the use of striping. Striping breaks up the video data into smaller chunks and divides those chunks across all the physical drives in the array. RAID 0 offers speedy reads and writes, but no redundancy whatsoever. If you lose one drive, you lose all video from all the drives.
RAID 1 — Providing a mirroring capability, this RAID level takes a single drive and completely duplicates it on another drive of the same size. If one drive fails, the other is ready to take over and nothing is lost. RAID 1, however, does not give you any read/write performance boosts.
RAID 5 — Referred to as striping with parity, RAID 5 is the standard in our industry. Just like RAID 0, the data is broken into pieces and written across all the drives. In this case, however, an extra piece of information is added to the mix, called parity. The parity information is stored in the form of a mathematical calculation and contains data from all the other drives. If any one of those other drives fails, the parity data can be used to rebuild a new replacement drive. Therefore, no video is lost.
One drawback to RAID 5, though, is that if a second drive fails before the first failed drive is rebuilt completely, then all video is lost across all drives. Of course, the odds of that happening are slim, but it does happen.
RAID 6 — Recently, RAID 6 has burst onto the scene and is becoming very popular. This version is very similar to RAID 5, but with a second parity block written to each disk. That second block gives the array the ability to stay running if two drives go down together. RAID 6 is rapidly becoming the standard to beat.
Next month we’ll examine some of the connectivity and network options for creating large video storage systems.
MCSE- and CCNA-certified Steve Payne has more than 15 years of industry experience and heads Convergence Consulting, an IP and security solutions consulting firm. Be sure to also read his Integrated Thoughts blog.