Alejandro Freixes
Mar 12, 2012
Featured

Beyond Anobit: Apple's game of thrones forces rivals to kneel

CEO Tim Cook speaks during the unveiling of its iPad 3Apple has an insatiable appetite for market control. Every acquisition, every strategy shift and every product demands analysis from every conceivable angle. While industry analysts have already spouted the obvious conclusions regarding Anobit’s acquisition and the buzz has entered a low burn phase, there are a lot of blanks yet to be filled. Anobit’s acquisition by Apple in December deserves a more context-rich look, especially in light of Apple’s rearrangement of Anobit management and the departure of Ehud Weinstein, the cofounder, chairperson and CEO, as well as Apple’s aggressive recruitment of engineers for an Israeli R&D center. In order to better understand where Apple is taking Anobit’s intellectual capital and the SSD flash memory market in general, as well as how industry dynamics with competitors like Samsung, Micron, SanDisk, SandForce and STEC will shift, it is helpful to understand the origins of Anobit and the bricks that laid its foundation. Patexia will examine the founding intellectual property framework of Anobit, the current state of the SSD market and the ramifications of Anobit’s acquisition for the electronic game of thrones.

The Mind and the Patents Behind the Curtain

Leigh Engineering Faculty Boulevard, Tel Aviv UniversityEhud Weinstein, a Professor at the School of Electrical Engineering at Tel-Aviv University, is also a Fellow of the Institute of Electrical and Electronics Engineers (IEEE). His credentials are extensive, including the publication of over 50 papers in leading scientific journals and his accomplishments as a Research Affiliate at the Research Laboratory of Electronics at Massachusetts Institute of Technology (MIT). He has served as cofounder, chairman and CEO of Libit Signal Processing, which was acquired by Texas Instruments in 1999, and he received the TI Senior Fellow in recognition of his contributions to Statistical Signal Processing.

The first several patents in which Ehud Weinstein was listed as an inventor were the result of his work with MIT and would later lay the foundations for his ventures into flash memory control with Anobit. This multi-channel signal separation patent from 1993 is listed via citation for the multi-channel signal separation using cross-polyspectra patent that would eventually lay the seeds for Anobit’s inception. Now, Anobit has 65 patents and publications -- 41 pending patents and 24 patents as of March 12, 2012. By analyzing the issue date of granted patents, we can observe that there was a sharp increase in 2011. Some of their most recent granted patents are “Wear level estimation in analog memory cells” (December 27, 2011), “Reading analog memory cells using built-in multi-threshold commands” (November 29, 2011), “Memory device with multiple-accuracy read commands” (November 15, 2011), and “Estimation of non-linear distortion in memory devices” (November 15, 2011).

Flash Memory is Kind of a Big Deal

In February 2007, Anobit Technologies was founded as a fabless semiconductor start-up focused on improving the growing NAND flash memory market that has gone from grossing $8.2 billion in 2001 to over $20 billion in 2010, and is expected to reach $26 billion in 2012. NAND-based flash memory is utilized in solid-state drives (SSDs), which are fast becoming a hot commodity in 2012. Unlike traditional magnetic disk storage like hard disk drives (HDDs) that use movable read/write heads, SSDs use microchips that are less susceptible to shock than the electromechanical HDDs and can even retain memory without power due to NAND technology. They offer lower access time and latency and are silent, but cost more per gigabyte to create. Depending on the number of parallel NAND chips used in a device, bandwidth can be scaled.

A Micron Technology technician works with equipment used in semiconductor manufacturing (Photo courtesy of Micron Technology/Handout)While SSDs began being implemented as early as the 1970s and 1980s in semiconductor memory for early supercomputers, they have evolved significantly over the years. Most recently, controllers that bridge NAND memory components and host computers were improved when Intel and Micron created ultra-fast SSDs with data striping and architectural interleaving in 2009. Two years later, SandForce released consumer-grade SATA 6 Gbit/s SSD controllers that compress data with controller chips prior to sending it to NAND chips -- this netted controllers that could support 500 MB/s read/write speeds.

SSD is not an easy market to predict, besides the obvious forecast that it will continue being highly lucrative. This is largely due to the fact that SSD can come in many different form factors -- like PCIe, rackmount and USB -- as well as various memory types -- like RAM SSD, SLC flash SSD, skinny flash SSD, etc. -- each of which serve a specific function and come with both advantages and disadvantages. Add to this the intellectual capital behind each iteration of SSD -- be they software gurus or chip afficionados -- and the subtlety of each evolution varies anywhere from nuanced controller evolutions to substantial architectural makeovers. Fusion-io (FIO), for example, is known for its software-based expertise and is closely followed as a prophetic force by industry commentators, especially in light of its successes with PCIe SSD -- for example, using its latency reducing Auto Commit Memory (ACM) extension, Fusion-io exceeded one billion input and output operations per second (IOPS) in a recent tech demo. One of the wizards behind the Fusion-io curtain, Donald Basile, was snatched up by Violin Memory in 2009, a company that is also a pioneer in the PCIe SSD market. The cross-pollination of brains in the memory industry is important to watch. Just tie the right mental threads of enough industry players together and the exponential return can be incredible.

Given that the latest breakthrough in flash can come from just about any conceivable angle, be it a technique or an invention, it is worth keeping an eye on the leaderboard for the fastest SSD in the market.

Why Anobit Threatens the Competition

So, in such a complex web of industry power brokers and kingmakers, why is Anobit so special as to warrant the almighty Apple’s attention and vast reserves of gold? Simply this -- Anobit departs from the typical fiddling with architecture and manufacturing by focusing its efforts on predicting errors in flash-memory with their Memory Signal Processing (MSP) tech. These signal processing algorithms increase speed, endurance and lower system costs by utilizing a variety cell (MLC) NAND, as well as MLC endurance and performance with triple-level cell (TLC) NAND, thus reducing the cost per-bit.

Samsung Group headquarters at Samsung Town, SeoulApple has taken a keen interest in Anobit for two simple reasons on the surface --  tech and talent: flash memory controllers are a key component in Apple’s iPads, iPhones and various other products and Anobit’s chip engineers are premium grade intellectual capital. The acquisition, however, has much greater significance for competitors, especially Samsung. The reason this stings Samsung is twofold: First of all, Samsung had a large contract with Anobit for its TLC NAND, which means Apple can now leverage their possession of Anobit by cutting off access to the crucial NAND flash in its main competitor’s products. Secondly, because Samsung itself was a supplier of chips for Apple, they have lost valuable business.

Other customers of Anobit, like Micron and Hynix, are now forced to look elsewhere for their supplies. Micron and Intel together have created IM Flash Technologies, where they design and produce NAND flash memory. While Micron is best known for its production of semiconductor devices, they are also investors -- in November of 2010, Micron’s first investment was Anobit, as part of a $32 million fundraising effort led by Intel Capital. Micron Ventures managing director, Matt Freeman, told Globes that, "I certainly regret that we did not acquire Anobit before Apple did … It could have been a good option for us. You can always understand the true value of a company when there is talk in the hallway about the significance of buying it. After Anobit was acquired, we were asking ourselves if this was bad for us." It disrupted Micron’s plans for entering the burgeoning and highly lucrative enterprise data storage systems market, which even led to them asking Apple if they planned on using Anobit's enterprise talent pool. They were told that Apple would not let any of the Anobit chip engineers go and that the team would be integrated into its new data center. Could it be that, among the many feathers Apple is placing in its hat with Anobit, it also wishes to enter the enterprise market?

SanDisk, the current leader in NAND controllers, is also concerned. Unlike Samsung and Toshiba, SanDisk does not supply raw NAND flash memory. Like Anobit, it offers flash memory storage solutions. Like Micron, Sandisk was often flirting with the idea of acquiring Anobit, but they opted to acquire Pliant instead. While Samsung has been left out to dry, because they are unable to deliver TLC in volume due to an inability to integrate chip level expertise with signal processing, SanDisk has managed to combine the two. Anobit’s own CTO, Avraham Meir, was CTO at M-Systems -- acquired by SanDisk in 2006 -- and Vice President Corporate Engineering at SanDisk prior to joining Anobit. Given Avraham Meir’s NAND flash technology expertise, many of SanDisk’s trade secrets were no doubt incorporated into Anobit’s tech. SanDisk’s TLC ingenuity is likely intact, however, and Apple lists it as one of the top suppliers in its list of 156 companies from which 97% of their materials and manufacturing spending derive.

Who’s Running the Show at Anobit Now?

Matam hi-tech park in Haifa, IsraelIn terms of management, Apple CEO Tim Cook revealed in the Q1 2012 earnings call that Apple Senior Vice President of Hardware Engineering Bob Mansfield will be in charge of the Anobit team; they are also going to be integrating the Anobit team into Apple’s current workflow. If Anobit ends up as the starting point for Apple’s Israeli R&D center, then they would likely continue staying put in Israel. However, if they get sent to Apple’s headquarters in Cupertino, then the R&D center will likely develop with less of a central focus on Anobit. Either way, Anobit’s executives will not be leading the center in Israel, as Apple Vice President Johnny Srouji -- who was a senior manager at Intel’s Israel-based R&D firm before joining Apple -- will be heading the team. While Bob Mansfield is still primarily in charge, Johnny Srouji is expected to take a more active role in leading the team -- he will, of course, still report to Bob Mansfield. Former Texas Instruments CTO, Etai Zaltsman, will also serve as a high-level employee in Apple’s Israel offices, with CEO Ehud Weinstein departing after ensuring a smooth transition.

Let’s Put the Pieces Together

As far as what Apple plans to do with its $400-500 million purchase of Anobit to incorporate it into future plans? Given the planned Israeli R&D center for Apple, there is speculation that Anobit could very well be the foundation for that center. This center will be rubbing shoulders with the three main local software giants in Matam hi-tech park in southern Haifa, Israel -- Intel, Microsoft and IBM. Helming Apple’s foray is Aharon Aharon, a veteran Israeli whose high-tech ventures include Camero Tech Ltd., which focused on developing Radio Frequency (RF) based imaging systems -- specifically, he will be in charge of running the center. In addition to breaking into the Israel high-tech scene, Apple has positioned itself globally as a provider of NAND-based flash memory solutions.

The effects on Samsung, Micron, SanDisk and other competitors is profound, not to mention the money saved by developing in-house rather than relying on outside suppliers. More than anything, the raw brain power of the engineers will serve Apple well, whether it decides to branch out into the enterprise SSD market or just continue polishing its current offering of products. In light of Apple’s sometimes predatory litigation and business practices -- one need only look at its recent browbeating of Proview in China, the arguably needless crushing of Kodak in its final moments and how it even tried claiming ownership of something as simple and broad as e-commerce -- it is easy to see how much of a stranglehold Apple will apply to the flash memory market. When you are the most financially successful company in the world by more than a $100 billion lead, why relax your control, influence and intimidation? Apple, like any king or military general, is playing a very long and calculating game. Consider how Apple turned to Japanese suppliers Toshiba and Elpida Memory for its NAND Flash needs in September, 2011 -- could Anobit’s firepower be leveraged to bring Japanese memory suppliers in line?

In the short-term, while Anobit is both an entry ticket into the Israeli scene and a choke hold on NAND-based flash memory solutions, in the long-term we should be keeping a very close eye on how Apple wields the elite engineering troops it has begun assimilating into its ranks -- what uncharted realms will they conquer? While the Anobit acquisition may seem fodder for yesterday’s headlines, the corporate game of thrones makes its boldest moves in the seamless weaving of threads behind the curtain. Anobit’s engineers, like chess pieces, will be moved around, perhaps to departments unfamiliar even to them. Then, months or even years from now, their contributions will change the tech game -- but by then, the media may have lost track of the cards that Apple so deftly reshuffled into its enigmatic deck.Apple CEO Tim Cook, during the iPad 3 unveiling in San Francisco on March 7, 2012, shows how many personal computer units the competition ships in comparison to the iPad