By now everyone has seen the great marketing adverts from AT&T where a interviewer is sitting in a classroom asking some younger students basic questions like, “Would like you like more or less? and my favorite “Which is better: fast or slow?”. Of course all of the kids scream FAST! Even to the point where one little girl talks about being slow and then being turned into a werewolf and then crying because all it wants is to be human again… Then the tagline, “It’s not complicated”.
No matter if we are talking about mobile phone, sandwich deliver (Jimmy John’s ROCKS!) or flash based areas, the “how fast can you …” is real. We can time a person delivering food, we can meter the signal on a mobile phone and we can measure the response time of a flash array. But what is more important than being fast? Simplicity.
There are many new start ups entering the flash storage market and some of the larger vendors are snapping up these little guys like little bunny foo foo (Not sure about the bopping part).
What I have seen is most of the solutions sound good on paper and power point but once in the sandbox they are complicated and take weeks of tuning to get the performance as advertised. There seems to be a growing trend in the flash market of “just get it in there” and then deal with tuning later mentality. This is not surprising from where its coming as they typically have more engineers on a customer to make sure their legacy gear behaves properly.
The IBM FlashSystems 710/720 and 810/820 were announced last week with a huge kick off meeting in New York and a huge social media push which I thought was marvelous. There was announcements of IBM setting $1 Billion (Yes, Dr. Evil does now work at Almadean Labs) for research and product enhancement of Flash technology. I think someone actually said something on our tweet chat that day like “Who else in the world could spend as much money on one particular part of hardware but IBM?” and I think the answer is fairly clear, no one.
What will come from this huge R&D investment you may ask? I suspect IBM is looking at how to take the technology and sprinkle it through out the storage portfolio and into server side as well. There are already some vendors who are getting started with adding flash PCIe cards into the hosts but they are typically storage companies trying to talk to server people. Not saying server people are different than storage people, but its a different language at times. Again the question comes up “Who else but IBM can talk servers, switches, storage and software and offer good pricing and support?”; no one.
I suspect the neigh sayers will be ramping up with their FUD about the IBM FlashSystem and how its this and not that, but let me tell you:
- It is fast
- It is easy to manage
- It is efficient
One aspect that I see as HUGE win is having the FlashSystem behind our SAN Volume Controller (SVC) and allowing the Easy Tier program figure out what data needs to be on flash and leave the rest on old rusting spinning drives. This becomes very interesting because now you get better performance out of your storage and increase the usability of the FlashSystem by always keeping the hot data on the fastest device. Per the IBM SSIC the FlashSystems are supported as a device that SVC can virtualize with different OS types.
I believe this is a very interesting time to be in storage as we are seeing the change in storage designs. I hope the new flash systems can be adopted by main stream customers and helps drive the cost down even further so we can start looking at solutions like Flash and Tape only. Needless to say there are a lot of ideas and cool things coming out of IBM Labs and Flash is going to be one of the biggest.
I have spent most of the day looking over the products from TMS (Texas Memory Systems) that IBM just acquired. One of the questions I have always wondered is how to map performance back to a certain technology. When dealing with DRAM and Flash devices there seems to trade offs on each. The first that comes to mind is DRAM requires some sort of battery backup as it will loose the data contents when power is lost. Most DRAM cards and appliances have this under control with some sort of destaging to SSD or they have some sort of battery attached to the IO card that allows the DRAM time to hold information until power is restored.
DRAM is typically faster than its flash cousin as well as reliable and more durable. Typically there is less controller latency due to the lack of complexity of wear leveling and garbage collection. DRAM is still more expensive that Flash and has the problem of needing power all the time.
When looking at systems to find out how to decide which solution fits your environment it comes down to price and IO. The DRAM solutions are usually smaller in size but can push more IO. For Example the TMS 420 is 256GB of storage in a 4U frame that can push 600,000 IOPs. Not bad if you need 256GB of really fast space. This could be used for very high transaction volumes. This can be deployed with traditional storage and used for the frequently used database tables and indexes while lower IO tables can be thrown over to the spinning hard disk side.
In comparison the TMS 820 Flash array delivers a whopping 24TB in a 1U space and can push a meek 450,000 IOPS. This is somewhat incredible as the footprint is small and dense but still gives you the punch needed to beef up parts of your infrastructure. I stared running the numbers to compare this with say a V7000 with all SSD drives and we can’t come close. You could virtualize the system under the SVC code (included in the V7000) and use the Easy Tier function to move hot data to and from the TMS array which gives you the performance needed. I see why IBM decided to acquire TMS now.
So who wins in a DRAM and Flash discussion? The vendors of course, they are the ones who are going after this market aggressively. I think most consumers are trying to figure out if its needed to spend money on moving a database from 1500 disks at sub 20 ms response to using 200 larger disk and adding the DRAM or Flash device to keep the same latency. As an architect I want to keep in mind how much space and environmentals all of those disk eat up and having an alternative even if it costs more up front, is appealing.
Yesterday the Education Secretary Arne Duncan called for the nation to move as fast as possible away from printed textbooks and toward digital ones. “Over the next few years, textbooks should be obsolete” he claimed. Apparently the Secretary who has never taught a child in his life has seen how other countries like South Korea have moved to a digital text instead of traditional books. While I do not disagree that digital readers are becoming more of the norm and even my son is really into reading books on the tablet, I don’t understand why children who don’t like to read now will become avid readers because its on a screen vs a paper text.
I think the other issue is money. Most states are now trying to keep the roof from leaking and feeding kids who don’t get enough to eat at home. I struggle to comprehend the ability of each state to provide millions of students with some type of e-reader and carry the insurance needed to repair/replace the devices when kids spill, drop, break those devices.
From a storage sales standpoint would I love if each school system had their own copies of the 100 page Algebra II book I had to carry around in my backpack on a IBM storage array. Maybe there is a market for some startup to create a reader that is cost effective, reliable, easy to maintain and kept up to date. You could have sheet music on it, supplemental videos and help for kids that may need that extra help when mom and dad can’t be bothered.
I guess when you look at the overall idea of helping kids learn more with better tools then why not use technology? But I dont think we are going to improve our test scores by having kids learn from ipads. We still have to get kids to want to learn and it starts at home with the parents and at school with teachers.
IBM Storage has been going very strong with the introduction of the V7000 Unified into the marketplace. Combining the V7000 disk array with the SONAS code and hardware, and out comes a box that can provide customers with direct SAN attachment and provide NAS to his demanding shares. This system comes in three boxes and has all the parts and pieces you need to setup the system in a jiffy. But before you go opening boxes and plugging in cables there are some things you need to plan for before racking and stacking.
The first thing is to prepare for the installation by filling in the planing guides found on the Info Center. There you will find guides on hardware, software, network and storage configuration planning. The Info Center has more information and is the central place to find all documents about administration, monitoring and troubleshooting for the V7000 Unified. Other systems like XiV, SONAS, SVC, N series all have Info Centers, just do a Google search and you can find them quickly.
V7000 Unified is designed so anyone can setup the system with little to no training. The install starts with a USB key that has a little wizard that asks you questions about the configuration of the system. This wizard then saves a file that is used to program the nodes/canisters with out you having to connect with a cable and answer 100 questions on a Putty session.
Now that you watched the video and have read all of the documentation, you are going forward with your install. As easy as it is to install a USB key sometimes things go wrong. IBM has put together a document that goes into a fair amount of detail some of the most common errors that people are running into during installations. Other information like links to the quick install guide and the support page will help you in the future find resolution. If you fat finger a IP address or pull the USB key out too soon, this document will guide you through the steps on correcting the ship and put you back on course.
The 5 page document can be found here along with other information about the V7000 Unified