The discussion on which is the better choice, Fiber Channel (FC) or iSCSI, will get different answers from different people.

It is certainly an arguable case that Fiber Channel is more expensive, but the case can also be made that it has become less expensive over the years.

With the masses moving into the 10Gb iSCSI market, the cost differences is not so great anymore.

So is FC dead? It is probably as dead as the mainframe computer.

Many have predicted that mainframes will die over 20 years ago, but the mainframe is very much alive and thriving today. In fact, the server virtualization world like VMware, Microsoft Hyper-V and Citrix Xen hypervisors are foundationally based upon mainframe technologies.

FC has been and will remain a solid contender as a transport for data. It operates on a lower stack than iSCSI. From a foundational level, FC inherently has an advantage.

iSCSI has come a long way since it first came out. With many enhancements like Jumbo Frames, and DCB (Data Center Bridging) iSCSI is a formidable solution, even with the overhead of the IP stack. With 10Gb iSCSI going mainstream competing with 8Gb FC, the lines are blurred with differences in some pros and cons. Becoming more of an argument of preferences and what each establishment sees as an advantage.

So is FC going to die?
Not in the near future, for there are enough out there that are “old school” like me. Don’t get me wrong, I am a big proponent of iSCSI, but there are just some applications that I will reserve just for FC. With the next release of FC going to 16Gb and iSCSI going to 40Gb, knowing your application should determine your transport type.

Other than that, it’s really a question of, “Will it be Chocolate or Vanilla?”

As you may remember when SATA drive technology came around several years ago, it was a very exciting time. This new low cost, high-capacity, commodity disk drive revolutionized the home computer data storage needs.

This fueled the age of the digital explosion. Digital photos and media quickly filled hard drives around the world and affordably. This digital explosion propelled companies like Apple and Google into the hundreds of billions in revenue. This also propelled the explosive data growth in the enterprise.

The SAN industry scrambled to meet this demand. SAN vendors such as EMC, NetApp and others saw the opportunity to move into a new market using these same affordable high-capacity drives to quench the thirst for storage.

The concept of using SATA drives in a SAN went mainstream. Companies that once could not afford a SAN can now buy a SAN with larger capacities for a fraction of the cost of a traditional SAN. This was so popular that companies bought SATA based SANs by the bulk, often in multiple batches at a time.

As time progressed, these drives started failing. SATA was known for their low MTBF (mean time before failure) rates. SATA SANs employed RAID 5 at first to provide protection for a single drive failure, but not for dual drive failure.

As companies started to employ RAID 6 technology dual drive failure protection would not result in data loss.

The “Perfect Storm” even with RAID 6 protection looks like this…

- Higher Capacity Drives = longer rebuild times: The industry has released 3TB drives. Depending on SAN vendor, this will vary. I have seen 6 days for a rebuild of a 2TB drive

- Denser Array Footprint = increased heat and vibrations: Dramatically reducing MTBF

- Outsourced drive manufacturing to third world countries = increase rate in drive failures particularly in batches or series: Quality control and management is lacking in outsourced facilities resulting in mass defects

- Common MTBF in Mass Numbers = drives will fail around the same time: This is a statistical game. For example, a 3% failure rate for a SAN in a datacenter is acceptable, but when there are mass quantities of these drives, 3% will approach and/or exceed the fault tolerant of RAID

Virtualized Storage = Complexity in recovery : Most SAN vendors now have virtualized storage, but recovery will vary depending on how they do their virtualization

- Media Errors on Drives = Failure to successfully rebuild RAID volumes: The larger the drive the chance of media errors become greater. Media errors are errors that are on the drive that renders small bits of data to be unreadable. Rebuild of RAID volumes may be compromised or failed due to these errors.

Don’t be fooled into having a false sense of security but having just RAID 6. Employ good backups and data replication as an extension of a good business continuity or disaster recovery plan.

As the industry moves to different technologies other new and interesting anomalies will develop.

In technology, life is never a dull moment.

What is Cache(pronounced like cash)?

Cache is essentially memory that is used to enhance or to accelerate performance in a system. Cache is generally bolted onto storage devices to help facilitate bursts in data movement in a disk or a SAN system.

Cache does not, in any way, change the storage device but merely adds a “cushion” or “shock absorber”. The larger the cache, the more the system can absorb “spikes” and “peaks” in demands on the storage device.

The result… a better end user experience, but is that a good thing? Of course, but understanding what cache is and using it properly will result in optimal and proper use of such a tool. Ultimately, achieving a balanced system is what should be the goal of anyone using data storage today

Throwing Cache at Your Storage Problems

I have seen this happen all too often… Sounds cliché, but throwing “cache” at all your problems won’t make it go away. There is a point in any storage system where by adding a lot of cache will reach the point of “diminishing return”. This point of diminishing return will vary depending on the type or storage device configuration (speed of disks, raid level, connection type).

Cache can buffer up to a certain point, but at some point, the data still needs to land on the actual storage system. Adding cache is like using a credit card. It is great when shopping because it helps us feel better when and we get things so much easier and faster. Storage systems that have a lot of cache have to eventually pay for it on the backend. This cost comes in the form of write penalties (from RAID levels), IOPs support (lack of disk spindles or slow disks) and quasi-virtualized storage (typically a File System to emulate). The big problems come when the data that is being stored is used very heavily and consistently.

Cache reaches the point of diminishing return and shifts your problems to the backend. So remember, a good storage system will show its merits when it can boast the use of less cache. An example of this would be Compellent. Compellent’s truly virtualized storage can handle data issues upfront with little dependence from cache.

With the growing popularity of SSDs because of their robust speeds, some disk vendors like EMC is using them as a way to augment traditional cache. This will help, but essentially it is pushing the point of diminishing return further down the road. SSDs are also not the same as cache for it is great for read operations, but for write operations, it is reduced (as compared to read operations). Be careful of this tactic because the backend still needs to process the workload.

So before you go out and spend tons of cash, check and ask your storage vendor, how much cache is in their system and why. Don’t invest in a storage system that resembles our national debt… remember, we eventually have to pay in the end.

The recent flooding in Thailand has been the worse in decades. The people are now threatened with disease from the flood water and facing many other hardships.

The floods has also caused a hiccup in the world’s supply chain of hard drives. Thailand produces 40% of the world’s hard drives. With approximately 1000 factories closed from the floods, some of them making computer hard drives, companies will need to make contingency plans to obtain required parts for their businesses.

Affected industries include PC makers, the Auto makers, and other well known tech giants who are increasingly using storage such as Facebook, Google, and Amazon.

Analysts estimate that there will be 50 million fewer drives produced in the 4th quarter, or a -30% decrease in production.

Experts are predicting a substantial increase in the price of hard drives in the coming months as production is constrained.

Other than protecting your data, is there a difference in RAID levels? Definitely! Storing data is much like driving. Depending on what you are doing with your car or what the road conditions are, there are optimal vehicles for those situations. For example, if you have a family of 5, it would be ideal to drive a minivan, although you can get around in a Toyota Camry. If you are driving in the snow, it would be ideal to drive a jeep, although you can also use a BMW with traction control. Knowing the properties of RAID will help you architect the best storage system for your applications. In the world of RAID, there is something known as write penalty.

Here is the breakdown:

RAID 1 or 10 = Write Penalty of 2
RAID 5 or 50 = Write Penalty of 4
RAID 6 = Write Penalty of 6

To put these metrics in perspective;

RAID 1 or 10 is like a BMW M5, fast and expensive. It is the best for applications that demand high performance in both Reads and Writes. The drawback is that it takes 100% capacity overhead. So if you have 100GB of RAID 1 or 10 storage it will consume another 100GB of additional capacity. RAID 1 or 10 is ideal for Databases, Virtual Desktops or any application that demands high IOPS, particularly Write-intensive apps.

RAID 5 or 50 is like a Nissan Altima, multi-purpose and affordable. It provides for the most usable disk capacity while providing robust disk redundancy. The capacity overhead is around 30%, providing exceptional value. RAID 5 or 50 is the most widely used in IT environments today. In many cases, RAID 5 is used exclusively. In fact, most environments, with Database, Email and File servers use RAID 5 without issues. Just like the old saying, “the squeaky wheel gets the grease”, if the users don’t complain, IT departments won’t bother changing RAID levels.

RAID 6 is like the Minivan. Made for capacity and not for speed. It is typically used for 1TB drives and above. RAID 6 is designed to withstand double disk failures. When a 1TB drive (or larger) fails, it takes a lot longer for the replacement drive to rebuild (I have seen around 15 – 24+ hours). Generally, the longer the rebuilding takes, the larger the window for another drive to fail, hence, exposure to data loss. So RAID 6 is good for this purpose as it provides ample protection for drives (usually 1TB or higher) that have lower MTBF (Mean Time Before Failure). The drawback with RAID 6 is the write penalty is high. This is not a good RAID level to use for Write-intensive applications. RAID 6 is usually used for data archiving, data warehousing and storing static data (i.e., PDF, JPG, etc).

Although there are considerable write penalties for RAID 5 or 50 and 6, the Read penalty is the same as RAID 1 or 10. The benefit of a Read intensive environment is that RAID 5 or 50 and 6 can be used without the high cost of RAID 10. So before you choose your RAID level, take a look at what you are doing with the data that will reside on it. You will thank me later for that…

In computer terms, “to boot” means to start or to begin. To re-boot is to re-start or to start over. As we all know from calling tech support to help us with a problem, we hearing it time and time again, “First, have you re-booted your computer?” And sure enough, it fixes about more than half of our computer problems. In fact, we don’t call support anymore without rebooting first.

Oh how loosely we use the term “reboot” these days…for example, “rebooting” a movie, ie. Spiderman, Karate Kid, Matrix etc. How about rebooting relationships? (What is the success of that one?) Friendships? While rebooting is good for computers, I can’t seem to agree with “rebooting” other things like Companies. This one really baffles me, HP. I don’t know what happened? Does anyone?

HP managed to self destruct themselves in a matter of days. First by following in the footsteps of IBM, HP is selling their PC division. I can see how this can work for IBM, but HP is clearly not IBM. Then after spending a ton of cash on developing a $500+ iPad competitor only to “dump” their tablets like it was toxic. It went from $500+ to $100 overnight. If you are one of the lucky few to get your hands on one these, that’s awesome! In an effort to “reboot” HP, they replaced their CEO Leo Apotheker with Meg Whitman the former CEO of e-bay. I think that is a move that will stop the bleeding, but it will be interesting to see what happens in the long run. Technology is such an unpredictable industry.

I honestly like HP and they make good, reliable and high-value products. But I sense that HP has lost its identity and was trying to be someone they are not. You can’t do something half-hearted and expect to succeed. Look at Apple. A company that stand behind their products and decisions. They live or die by it. They went from being a company that almost went bankrupt a few times to a company today with more cash than the US government. So I hope HP can successfully “reboot” themselves and put their hearts back into their products as they are a true American household name. I have seen too many good companies “reboot” themselves into history.

Have you ever hear of the term “Defrag” and wondered what that meant? Defrag is actually “de-fragment” or the act of un-fragmenting. To defragment implies that something is fragmented. In the context of disk drives, fragmentation is normal. Fragmentation is a by-product of storing data and it happens over time. As data is constantly written on hard drives, data is also constantly deleted.

Here is an example that may help you understand what data fragmentation is. Think about parking spaces; when you go to a theme park with friends (when the park first opens), it is likely that you will park together. It is nice because you are all together. Then you all head out to a lunch spot outside of the park. When you all return, it is unlikely that you will be parked together. That makes meeting up and coordinating activities harder.

That is the same with data, over time, data becomes fragmented. Data is spread out to different parts of the disk. Data is broken up and shoved open data spots where other data once lived. When data is fragmented accessing data becomes slower because the disk head needs to move back and forth more to get data.

Fortunately, there is a utility built into Windows that will help you to clean up fragmentation; it is called Defragmenter. Best way to access that in Windows 7 is to go to “Start” and type in “disk defragmenter”. Then run the utility. Defragmentation is a tedious process and should be performed when the computer is not in use. Depending on how big your hard drive is and how much data you have, it can take a long time. Nevertheless, it is logical housekeeping and should be performed to keep your computer operating in tip-top shape.

If you have been around technology in the 80′s you may know that Megabytes=Power.
2k 8k, 16k etc… was what were upgrade options for RAM for my computer.

Floppy disks were the mass storage option and it started out as an 8 inch floppy. The 5.25 inch floppy arrived and was much smaller than the 8 inch disk. When the 3.5 inch disk came out, it was like heaven! Not only was it small, it stored 1.44MB of data. I bought a 200MB hard drive and paid $400 bucks for it. I had bragging rights for a long time! Like any guy who ponders about the next level, Gigabytes were untouchable although 1 Gigabyte had drives were around, but they were very expensive! I saw one at that time for $30,000! I can buy a well equipped Mini Cooper with that. After the Gigabyte era came the Terabyte drives. Since we are now caught up and currently at 3TB drives, what’s next? How about after that? Well, here is the breakdown.

8 bits = 1 Byte
1024 Bytes = 1 Kilobyte
1024 Kilobytes = 1 Megabyte
1024 Megabytes = 1 Gigabyte
1024 Gigabytes = 1 Terabyte
1024 Terabytes = 1 Petabyte
1024 Petabytes = 1 Exabyte
1024 Exabyte = 1 Zettabyte
1024 zettabyte = 1 Yottabyte

Apparently, there isn’t anything that has been officially named after Yottabyte yet, but I am sure we will get there soon enough. To think that I have 32GB in my cell phone that I throw around and to know that a 1GB hard drive went for $30,000, my cell phone could have been worth over $960,000 bucks! Can’t wait to have a Yottabyte on my keychain…

Do you know we live in amazing times? When I was growing up and if I wanted to learn to dance, I would have to take lessons. If I wanted to learn construction, I would get an entry level job as a construction worker. Today, you can Google virtually anything and learn almost anything from the Internet. As I was thinking today about how far we have come with storing data. I wanted to take this time to simplify this nifty technology so that many can “Google” RAID technology and understand it in about 5 minutes.

Although RAID has been around for a long time, most people who are not in IT won’t know what RAID is. There has been some consumer version of hardware RAID cards for the home, but is not commonly used. Let’s first start with what RAID stands for; “Redundant Array of Independent Disks”. When RAID was first introduced, it stood for “Redundant Array of Inexpensive Disks”. The acronym was changed to reflect the changing nature of hard drives and RAID sets. Basically, RAID is a data protection method that employs different data storing algorithms using a set of disks. There are different levels of RAID which is designated by a number following the term “RAID X” (X being the RAID level). I will break down the different RAID levels for you.

RAID 0
This RAID level is striping without parity. Striping is the ability to store data across multiple drives. Parity is an error correction method that is used in RAID and is a core mechanism in rebuilding failed drives. This RAID level offers no protection… Yes, this is the only RAID level you probably don’t want to use. The only advantage of this RAID level is increased capacity and throughput because there are more disk spindles in the RAID disk set. The minimum RAID 0 set contains 2 drives. Some external home use drives use RAID 0 to increase capacity in which I DO NOT recommend unless you have another set of backups somewhere else and capacity is paramount. In my book, capacity never trumps reliability when it comes to data storage.

RAID 1
This RAID level known as disk mirroring (without parity). Simply put, RAID 1 is a duplicate image of the main disk on another disk. This is also called duplexing. This RAID level is usually done using hardware controllers, but can also be done using some operating systems that support disk mirroring or third part software. There are definitely advantages for using this RAID level, but is the most costly because essentially you will be buying double the usable disk capacity.

RAID 5
This RAID level is disk striping with distributed parity. What this means is that the data is distributed along with the parity data across all drives in the RAID set. This RAID level can tolerate a single drive failure which will need to be replaced. Upon replacement, the new drive must be rebuilt from the surviving drives. During this time, the failure of a second drive will result in data loss.

RAID 6
This RAID level is disk striping with double distributed parity. What this means is that this RAID set can tolerate 2 failed drives and still be operational. Failure of a third drive will result in data loss.

HOT SPARE
A hot spare is usually a disk that is powered on and spinning that sits in an array that does nothing but wait for a drive failure. The hot spare is then automatically rebuilt from the surviving disks in the RAID set . This will allow for minimized windows of exposure to data loss.

I will be expanding on RAID technologies in my future posts and hope this was helpful in understanding this complex but compelling technology.

The data storage market is becoming an interesting one. I remember about 10 years ago, we were all jokingly saying that we will all someday work for Microsoft because they had all the wealth and was putting everyone else out of business. We live in interesting times today where jobs are hard to find, but for a select few, companies are insanely in a bidding war for these people.

Who ever thought that Apple would be where there are today? For you younger folks that think Apple is “King”, they were paupers once… in fact, they almost went out of business a few times! I know, it is hard to believe, but my point is that companies are adapting to culture and trends to survive or thrive. Some do it better than others.

For example, when I grew up, furniture stores were boring places where we walk around and look at furniture. For kids, it is a torturous venture to say the least. Today, we have IKEA where we can go and drop off the kids in the ball pit and then enjoy a hearty family meal sipping lingonberry juice afterwards. We have come a long way in a very short time!

As for your technology, there are a few companies that want to be your one stop shop… Your Walmart of technology. 20 years ago, Dell was just a home computer company, now they have acquired Equallogic, Exanet, Ocarina and Compellent to compete with NetApp, EMC, IBM, HP etc…

Oracle was just a database company, but they bought Sun. EMC was a storage company but bought Legato, Isilon and VMware. HP was a server and printer company, but bought 3Par and Lefthand Networks. These are just a few small examples that left me scratching my head. I am sure many other people were also.

The bigger picture is there that there IS a strategy in all of these acquisitions. The strategy is to be your “One Stop Shop” for all your technology needs. Essentially to own the data center. This is interesting because as technology advances, unlikely unions become common place. One time rivals become united in marriage. The bottom line is that “we”, the consumer are the winners. Better pricing, better technology and responsive support are constantly being improved.

So as I check my news every morning, it is like a technology soap opera for geeks like me. So what’s next? Is Dell going to buy Brocade? Cisco to buy EMC? NetApp to buy Citrix? Apple to buy Dropbox? We shall see…