Looking Towards the Light

After a long hiatus because of school and personal matters, I am back in action and excited to talk about today’s topic. Before we get into that, here are a couple of announcements. First, I have finished my degree in Information and Telecommunication Technology, and currently re-working my resume to start sending out applications. Second, I am still working part-time for Greg Schulz, owner and founder of StorageIO, and you should check out his blogs [here].

 

As you might have guessed, today we are looking toward the light. No, not the desk lamp next to you, the sun outside (which you shouldn’t stare at by the way), or some metaphor. Today, we are going to be talking about optical storage. I originally found out about this medium through Greg Schulz, when we were conversing about different types of data storage. The conversation spread though the usual mediums, starting with solid state drives (SSD), moving to hard disk drives (HDD), and finally tape. Sometimes we would even shake things up and talk about all three at the same time *gasp* (details on those conversations will come on at a later time). At the end of the conversation, he asked me one question, “Have you heard of optical storage?”

 

I had studied optical networking before then, but had never even heard of optical storage. Naturally, my first questions where what is it used for, and if it was even in use today. He responded by asking me what I thought Blue-Ray discs were. He also mentioned Facebook doing something with optical storage if I was interested in learning more.

 

With that, I was on my way researching all I could about optical storage. I found a lot about what optical storage was, and how it works.

 

A magnified view of an optical disc with data burned on it. Credit given to igcseict.com

 

 

 

 

 

 

 

 

The picture above is of a magnified CD surface. Can you see the dots in the picture? The pattern of these dots is your data. When you are burning your favorite Justin Beiber song on a CD (if you still use them), your computer is creating a pattern on the CD to store that song. When inserted in a CD player, the laser inside will reflect the surface/pattern of the CD back to itself, and play the song.

 

Trying to find how it is used in the industry proved to be more challenging.

 

Facebook’s Blue-ray rack. Image: Kyle Owen

 

 

 

 

 

 

 

The short version:

-At this time all of Facebook’s cold storage is stored on HDD.

-They believe swapping to an optical system will be 50% cheaper then disk, and 80% more power efficient.

– Each disk is certified for 50 years of reliability

– Believes this is something better then disk, or tape for future demand for cold data.

 

If you would like the full article, you can click [here]

 

Is Facebook heading a paradigm shift in storage, or is their optical solution a dying hope to make optical a valuable alternative to tape archiving? My personal thoughts, and more will be explained in part two.

You can check out my other blog posts [here], and follow me on Twitter @StudentofIT. Until then, take care. 🙂

 

Cory Peden

 

Myth Busters-Tape Edition

Myth Busters

A while back my mentor Greg Schulz, founder of StorageIO, asked me what did I know about tape? I told him the topic has come up in my studies, but was looked down upon as a slow and old method for saving data. From this he asked if I would be interested in talking with someone who works with tape, and to shed some light on these myths I had. That is how I came in contact with Rich Gadomski. Rich works for Fujifilm as the Vice President of marketing for the commercial product division. During our talk, I learned my perspective of tape had been misguided. Below are some of the myths about tape, and the truth behind them.

Myth: Tape is slow

I hear this claim all the time with fellow students who jumped on the SSD bandwagon. The first question I ask them is why is tape slow? True if you are doing random reads tape will be slower than disk. However what if we want sequential reads, or sequential writes? Let us look at the specifications of Fujifilm’s most current magnetic tape LTO G6. Uncompressed LTO G6 has the ability to read/write sequential data at a rate of 160 MB/sec. If that data is compressed, the transfer rate can reach as high as 400 MB/sec. The issue is your network’s ability to feed data to your tape, not the tape itself.

Myth: Tape has low capacity

How many people who clam this really know the capacity of current tape? Before my interview with Rich I had no knowledge of the capacity of tape, but was surprised how comparable it was too Hard disk drives (HDD). I went to Seagate’s web site to learn current capacity of enterprise 2.5 and 3.5 HDD, compared it to Fujifilm’s LTO G6. The highest capacity of a 2.5 enterprise HDD capped out at  1TB, while a 3.5 enterprise HDD capped out at 6TB. LTO G6 currently has a capacity of 2.5 TB uncompressed and 6.25 compressed.

This is possible because of Barium Ferrite (BaFe). BaFe is a magnetic particle found in Fujifilm’s LTO G6 tapes. It is less dens then the metal particles found in older LTO tapes, and this allows more data to be stored on a single tape. Fujifilm predicts that this breakthrough will allow tape to reach a capacity of 35TB of uncompressed data in the near future.

Myth: Tape will be dead in 10 years

Let us look at what we have talked about so far, and see if this is true. We know tape is fast when it comes to sequential reads and writes. This makes tape perfect for recording, or recovering large sums of data.

Something Rich pointed out to me that I had never thought about is tape is cheap to maintain. You have to power SSD and HDD 24/7 costing money, and taking up resources. Yes this might not seem like much, but when you are powering hundreds of these devices, and powering the systems to keep them running (cooling, UPS systems, ect..) it starts to add up. Tape requires no power when it is not in use, and as long as it is stored within 60-95 degrees will maintain integrity.

Let us take a closer look at the integrity/reliability of tape, HDD, and SSD. It is not fair to compare the software that checks reliability since often times tape is used to save data after it has been on these other mediums for some time. This means the data has been checked multiple times if it is reliable before it is even written to tape. What we can look at is how all three compare when it comes to physical reliability.

When the tape in the cartridge snaps, is it possible to repair the tape, and recover most of the data? The answer is yes. The tape can be sent to a number of vendors to be repaired, and while there might be some data loss it is possible to recover most of the data. The same is true for HDD if it has been scratched, and SSD if a sector has gone bad. What does this mean? The answer is simply all three have roughly the same reliability physically.

We also know currently tape is competitive when it comes to capacity, and is only expected to grow in the future. The same is true for both HDD and SSD. While growth capacity for HDD has slowed down over the past couple of years new advances are currently being developed to ensure the continuant use of HDDs.

Closing thoughts

People need to stop seeing it as  tape vs. disk, tape vs. flash, or flash vs. disk.  My beliefs about tape, HDD, and SSD are the three should be working together to provide the best solution. Too often, we read about how one is going to replace the other, or how one is going to be gone in the next ten years. In reality, they have been coexisting for years, and will continue to do so for years to come. Next time you hear someone talking down about tape, educate them in their myths.

 

What’s Next

Over the next few weeks, a variety of topics are going to be covered. Optical, Buzzword Bingo, IPv6 are some of these topics.

I encourage fellow students of IT to leave suggestions on topics they would like to read about, and to engage in the conversation. After all, innovation cannot happen without conversation.

Feel free to follow me on twitter at Studentof_IT

Until next time

Cory Peden

 

Data Networking: Physical vs. Protocol

Now that you have learned a little about myself from my about me post, let’s dive into our first topic in my on-going series as a student of IT (that’s also my twitter handle btw @StudentOf_IT).

Physical and Protocol Networking

I believe students often forget with networking the physical medium, and the protocol used to transport the data are independent from each other. For example, my education curriculum or program focuses mostly on Transmission Control Protocol (TCP) Internet Protocol (IP) over Ethernet. Although, there is a relationship between the two, and both a physical medium and protocol are required to transport any data, IP is not reliant on Ethernet, and Ethernet is not reliant on IP.

Rules of the Road (or Network)

Confused yet? I was when I was trying to understand this. Picture the physical medium as a road. By itself, this road is meaningless, and dangerous without rules and control protocols. There is a clear path from one place to another, but the cars are driving at various speeds, sides of the road, and have little care for the cars around them. This is because a road by itself does not have rules. Have you ever been behind a group of people walking really slow taking up an entire hallway? This is because that hallway by itself has no rules. The people walking in the hallway make the rules or follow an established protocol.

Let us add the type of physical network cable into the equation now. For this example, we are going to use 1Gb Ethernet (e.g. 1GbE), however this applies to any type of optical or electrical cable (not to mention wireless). Ethernet is our rules on the road or physical medium. We have a speed limit of 1Gb (one billion bits per second), we must drive in given direction on a certain side (transmit or receive path) of the road, and any vehicle on the road must meet certain standards (Ethernet frame).

The next logical assumption is the protocol are the vehicles on the road right? Well half right, and this was the part I had the most trouble understanding. The vehicle itself is irrelevant. The cargo inside determines the protocol used. The protocol is providing you directions to how to reach your destination, and standards of how to handle the cargo.

Bring IT (The Network) Together

Let us put the two together now. The cargo will be using IP protocol, and the road will be Ethernet. You bought a shirt at Target that is too small for you. Because of this you want to return the shirt, and buy a new one. The shirt in this case is your cargo/IP packet (remember, this applies to any type of protocol). The shirt has determined you are going to Target, and because of the road (any physical medium), and the rules put in place on the road (ethernet) we know how to get there safely.

What if you do not have any cargo/requesting data?

This would be handled in the same way as the above example. The car might be empty now, but we were told how to handle the future cargo, and are already following the standards of how to transport it.

Wrap up

IP is not solely reliant on Ethernet, and Ethernet is not solely reliant on IP. In other words the road is not reliant on only one set of rules, and the rules can apply to more than one kind of road. If IP was implemented on a Fiber Channel network (as it was in the past), the road itself would be changed, but the rules would stay the same, like going from a paved to a gravel road. The rules on the road can be changed as well. Fiber Channel over Ethernet (FCoE) can be implemented on an Ethernet network instead of IP. This would change how we drive on the road. Britain driving on the left side of a paved road, instead of the right.

That is enough from me for now, check back soon to read my continuing series of experiences as a student of IT. 😛

Cory Peden – Student of IT

@StudentOf_IT

www.studentofit.com

 

About Me

Let us start with the basics. My name is Cory Peden, I am in school full-time at Century College pursuing an AAS degree in Information and Telecommunication Technology (ITT). After this semester I will be one class away from graduation. While going to school I work part-time in the food service industry, and am an intern with Greg Schulz at StorageIO.

IT was not my first career choice, or  second. When I first went to college back in 2007 I wanted to become a photographer, and travel around the world taking pictures of nature or architecture. After a year and a half of bad choices I was suspended. Lost not knowing what I wanted to do with my life I took a year and a half off. During this time I had become a certified trainer at work, and found I liked teaching people new things. This lead me to pursue a degree in education.

This lasted for about a year when I saw the direction education was taking, and all the red tape that goes behind being a  teacher. I did not want to go into a field that measured your success based on one standardized test, and not take into account conditions of the schools, or special needs some students might have. I also believe the whole education system needs a revamp, but that is a topic for another time.

Giving up on school, I tried my hand at management, but because of events happening in my personal life, I could not take the pressure, and was not mature enough for that much responsibility.

Everything turned around when one of my friends, who is an IT professional, showed me blade servers he was working on. He explained the concept of visualization to me, and how you could run multiple things on one single machine. This sparked my interest, and I started asking him questions. After a couple of hours of this he finally told me that I should enroll in Century’s IT program.

Two years later, I am now on track to graduate by the end of the summer term. Last semester, I was on the Dean’s list, I plan to take my CCNA certification test within 3 months of graduating. My end goal is to become a consultant, and/or run my own datacenter.

Looking forward to creating conversation.

Till next time

Cory Peden

BTW follow me on Twitter @Studentof_IT to be updated on the newest blog posts