RAID(s): Not Just for Vikings or Pillaging Hoards

For the purposes of our discussion today, we are going to ASSume (yea, I know what “assume” does) that you have found this article and will further read it because the word RAID is not only foreign but you're not even sure how to spell it. For those who are new to the concept of using computer storage systems, I'm especially talking to you folks who have been using DVRs and analog cameras for video surveillance installations, it turns out that RAID isn't a word after all. It's an acronym (the big word of the day, even though its not very “big”) that stands for Random Array of Incompetent Dunderheads – OK, I made that up but I'm sure we all have examples we can cite. Random Array of Independent Disks (R.A.I.D.) is not only pretty cool but its also extremely useful in systems requiring more than one hard drive and/or some form of redundancy. If you are an installation tech, a sales engineer, manager, VP (does VP really stand for Vacuous Pratt?) or someone just looking for a high level understanding of RAID then you should be able to get what you need right here. If, however, you are more inquisitive, a hard gore geek (“it takes one to know one” - and you thought I was a VP) or just want to dig in and learn the details that makes the technology work, there is a great write up in wikipedia.

RAID comes in multiple flavors, called levels. These levels can be nested (combined) to create new flavors. We are going to concern ourselves with the most common levels – i.e. what you are mostly likely to come across. These are levels 0, 1, 5 and 6.

RAID (Level) 0: This method combines all the of the disks in the array into one, possibly gigantic, hard drive (at least as far as the OS is concerned). The cool thing about this is that there is no loss in capacity as all drives are added together and since there is no math going on to handle redundancy the write speed is at maximum performance for a RAID. The down side of RAID 0, and this is a rather steep downside, is that there is no data protection as there is no redundancy. The upshot is that if you lose one drive in the array, you lose the whole array. And as each drive has a an independent probability of failure the possibility of a failure of the array increases as drives are added.

RAID (Level) 1: RAID 1 is often called mirroring as data is written to drives created a “mirrored pair”. Obviously this creates great redundancy and only minimally impacts read and write performance – compared to a single drive. The data in the array will remain viable as long as one of the drives is operating correctly. I'm sure you see the drawbacks: 1) the array's capacity is limited to a single drive and 2) you need as twice as many drives for the same capacity as single drives. As much as hard drive prices have come down over the years they are still not free.

RAID (Level) 5: For storage systems (SAN, NAS and DAS) used with NVRs, whether external or internal, this is the most popular arrangement. RAID 5 uses distributed parity (math) across all the drives in the array - minimum size is 3 drives. This gives you the ability to operate at reduced performance and recover from one drive failure which of course is highly desirable. But if a second one goes out your data is toast so you'll need to get failed drives replaced ASAP. And what is the price of this miracle you ask... One drive's worth of capacity and reduced write speed (compared to RAID 0). That means that if you have an array with 2TB drives it'll cost you a extra 2TB drive.

RAID (Level) 6: For the layperson, this is basically RAID 5 with even more drive failure tolerance. Upside – two drive failure tolerance. Downside – 2 drives lost in capacity and slower write speed. Few installations in the security world use this but for those applications where “failure is not an option” or the site is distant and you don't want to have to immediately run out to the job site RAID 6 may be worth the extra cost.

So, recapping for the VP's... RAID makes multiple drives look like one to the Operating System, comes in multiple guises , can provide data protection and is widely used with Network Video Recorders (you'll have to trust me on this one). There are some bug-a-boos with RAID, but that's a subject for another time.

Sing with me now (to the tune of War): RAID! Huh! What is it good for? Absolutely... quite a bit it turns out.

Start at the Beginning

A beginning is a very delicate time... OK, I'm a bit of a Sci-Fi Geek borrowing a line from Frank Herbert's Dune. But that aside, how does one start something new – a blog, a painting, a software program, overhauling a Chevy 350 small block, etc...

I figure that since this is my first blog no one is going to read it as nobody knows about it. That gives me the freedom to basically say whatever I want without ramifications. And yes I understand that this will remain on my permanent record but how many people will say, “Hey, I like this blog, I think I'll go back and read all the old issues”? I mean, the first part of the sentence is a bit of stretch to begin with. No, I think this is more like a preface than an introduction, even though there are introductory elements to the content, and let's face it not very many people (besides me of course) bother with the preface.

So with that rambling introduction out of the way let's get to the real introduction. My name is Graham Joys (clicking on my name will take you to my LinkedIn profile) and I am writing this blog on behalf of Vista Networking Solutions (VNS). We design and build Network Video Recorders (NVR) and you can learn more about us by clicking on our name.

While of course it is tempting to use this forum as a vehicle for company news and self promotion I'm pretty sure that that would become pretty boring after oh, say, one article. So it is my intent to create a platform to discuss technology within our Industry – namely physical security. Of course I reserve the right for a shameless plug at some point in the future (wink). OK now, you ask yourself, what the H E double hockey sticks does this mean. Well, I envision that in the future (when people are actually reading), we will present some tutorials, discuss industry trends and challenges, invite guest speakers, include a video or two and address topics championed by readers... all in a conversational manner with hopefully a little humor – at least enough to make you smile (or groan) if not chuckle but I draw the line at guffaws – those are just creepy. I should state up front that I am, like everyone else in business, biased; so the OpEd elements of this space will be slanted. I believe IP will (and actually is) take over our industry like it has everywhere else – if you don't believe this try talking to someone who used to work at Alcatel or Nortel. My goal here is to provide information that you will find interesting and useful. I am not competing with the industry journals, which are mostly filled with long articles by advertisers masking as case studies. Here we'll keep the content to one page so you can consume it easily and move on – sort of like a taco.

Alright then, I've made it through my first blog and already I've violated one of the tenants I stated above – I've not imparted any information, wisdom, message or anything else for that matter. If no wisdom is imparted and no one is there to receive it did anything really happen? So since I have a little space left and I'd rather not grapple with the philosophical implications of the previous question let me give a preview of the topics we are kicking around for future issues (in no particular order).

• RAID for Dummies (and me)
• The IP address
• The MAC address
• Video compression
• Power-over-Ethernet
• Market Segmentation
• Things to consider when...

I guess in hind sight the last paragraph was like one of the phantom endings in The Return of the King. This is really the end. If there was anyone reading please let me know – I suppose there are always hecklers.