I was asked some interesting questions recently by CEO & CIO, a Chinese business magazine. The questions ranged from how Chinese Internet giants like Alibaba, Baidu and Tencent differ from other customers and what leading technologies big Internet companies have created to questions about emerging technologies such as software-defined storage (SDS) and software-defined datacenters (SDDC) and changes in the ecosystem of datacenter hardware, software and service providers. These were great questions. Sometimes you need the press or someone outside the industry to ask a question that makes you step back and think about what’s going on.
I thought you might interested, so this blog, the first of a 3-part series covering the interview, shares details of the first two questions.
CEO & CIO: In recent years, Internet companies have built ultra large-scale datacenters. Compared with traditional enterprises, they also take the lead in developing datacenter technology. From an industry perspective, what are the three leading technologies of ultra large-scale Internet data centers in your opinion? Please describe them.
There are so many innovations and important contributions to the industry from these hyperscale datacenters in hardware, software and mechanical engineering. To choose three is difficult. While I would prefer to choose hardware innovations as their big ones, I would suggest the following as they have changed our world and our industry and are changing our hardware and businesses:
Autonomous behavior and orchestration
An architect at Microsoft once told me, “If we had to hire admins for our datacenter in a normal enterprise way, we would hire all the IT admins in the world, and still not have enough.” There are now around 1 million servers in Microsoft datacenters. Hyperscale datacenters have had to develop autonomous, self-managing, sometimes self-deploying datacenter infrastructure simply to expand. They are pioneering datacenter technology for scale – innovating, learning by trial and error, and evolving their practices to drive more work/$. Their practices are specialized but beginning to be emulated by the broader IT industry. OpenStack is the best example of how that specialized knowledge and capability is being packaged and deployed broadly in the industry. At LSI, we’re working with both hyperscale and orchestration solutions to make better autonomous infrastructure.
High availability at datacenter level vs. machine level
As systems get bigger they have more components, more modes of failure and they get more complex and expensive to maintain reliability. As storage is used more, and more aggressively, drives tend to fail. They are simply being used more. And yet there is continued pressure to reduce costs and complexity. By the time hyperscale datacenters had evolved to massive scale – 100’s of thousands of servers in multiple datacenters – they had created solutions for absolute reliability, even as individual systems got less expensive, less complex and much less reliable. This is what has enabled the very low cost structures of the cloud, and made it a reliable resource.
These solutions are well timed too, as more enterprise organizations need to maintain on-premises data across multiple datacenters with absolute reliability. The traditional view that a single server requires 99.999% reliability is giving way to a more pragmatic view of maintaining high reliability at the macro level – across the entire datacenter. This approach accepts the failure of individual systems and components even as it maintains data center level reliability. Of course – there are currently operational issues with this approach. LSI has been working with hyperscale datacenters and OEMs to engineer improved operational efficiency and resilience, and minimized impact of individual component failure, while still relying on the datacenter high-availability (HA) layer for reliability.
It’s such an overused term. It’s difficult to believe the term barely existed a few years ago. The gift of Hadoop® to the industry – an open source attempt to copy Google® MapReduce and Google File System – has truly changed our world unbelievably quickly. Today, Hadoop and the other big data applications enable search, analytics, advertising, peta-scale reliable file systems, genomics research and more – even services like Apple® Siri run on Hadoop. Big data has changed the concept of analytics from statistical sampling to analysis of all data. And it has already enabled breakthroughs and changes in research, where relationships and patterns are looked for empirically, rather than based on theories.
Overall, I think big data has been one of the most transformational technologies this century. Big data has changed the focus from compute to storage as the primary enabler in the datacenter. Our embedded hard disk controllers, SAS (Serial Attached SCSI) host bus adaptors and RAID controllers have been at the heart of this evolution. The next evolutionary step in big data is the broad adoption of graph analysis, which integrates the relationship of data, not just the data itself.
CEO & CIO: Due to cloud computing, mobile connectivity and big data, the traditional IT ecosystem or industrial chain is changing. What are the three most important changes in LSI’s current cooperation with the ecosystem chain? How does LSI see the changes in the various links of the traditional ecosystem chain? What new links are worth attention? Please give some examples.
Cloud computing and the explosion of data driven by mobile devices and media has and continues to change our industry and ecosystem contributors dramatically. It’s true the enterprise market (customers, OEMs, technology, applications and use cases) has been pretty stable for 10-20 years, but as cloud computing has become a significant portion of the server market, it has increasingly affected ecosystem suppliers like LSI.
Timing: It’s no longer enough to follow Intel’s ticktock product roadmap. Development cycles for datacenter solutions used to be 3 to 5 years. But these cycles are becoming shorter. Now, demand for solutions is closer to 6 months – forcing hardware vendors to plan and execute to far tighter development cycles. Hyperscale datacenters also need to be able to expand resources very quickly, as customer demand dictates. As a result they incorporate new architectures, solutions and specifications out of cycle with the traditional Intel roadmap changes. This has also disrupted the ecosystem.
End customers: Hyperscale datacenters now have purchasing power in the ecosystem, with single purchase orders sometimes amounting to 5% of the server market. While OEMs still are incredibly important, they are not driving large-scale deployments or innovating and evolving nearly as fast. The result is more hyperscale design-win opportunities for component or sub-system vendors if they offer something unique or a real solution to an important problem. This also may shift profit pools away from OEMs to strong, nimble technology solution innovators. It also has the potential to reduce overall profit pools for the whole ecosystem, which is a potential threat to innovation speed and re-investment.
New players: Traditionally, a few OEMs and ISVs globally have owned most of the datacenter market. However, the supply chain of the hyperscale cloud companies has changed that. Leading datacenters have architected, specified or even built (in Google’s case) their own infrastructure, though many large cloud datacenters have been equipped with hyperscale-specific systems from Dell and HP. But more and more systems built exactly to datacenter specifications are coming from suppliers like Quanta. Newer network suppliers like Arista have increased market share. Some new hyperscale solution vendors have emerged, like Nebula. And software has shifted to open source, sometimes supported for-pay by companies copying the Redhat® Linux model – companies like Cloudera, Mirantis or United Stack. Personally, I am still waiting for the first 3rd-party hardware service emulating a Linux support and service company to appear.
Open initiatives: Yes, we’ve seen Hadoop and its derivatives deployed everywhere now – even in traditional industries like oil and gas, pharmacology, genomics, etc. And we’ve seen the emergence of open-source alternatives to traditional databases being deployed, like Casandra. But now we’re seeing new initiatives like Open Compute and OpenStack. Sure these are helpful to hyperscale datacenters, but they are also enabling smaller companies and universities to deploy hyperscale-like infrastructure and get the same kind of automated control, efficiency and cost structures that hyperscale datacenters enjoy. (Of course they don’t get fully there on any front, but it’s a lot closer). This trend has the potential to hurt OEM and ISV business models and markets and establish new entrants – even as we see Quanta, TYAN, Foxconn, Wistron and others tentatively entering the broader market through these open initiatives.
New architectures and new algorithms: There is a clear movement toward pooled resources (or rack scale architecture, or disaggregated servers). Developing pooled resource solutions has become a partnership between core IP providers like Intel and LSI with the largest hyperscale datacenter architects. Traditionally new architectures were driven by OEMs, but that is not so true anymore. We are seeing new technologies emerge to enable these rack-scale architectures (RSA) – technologies like silicon photonics, pooled storage, software-defined networks (SDN), and we will soon see pooled main memory and new nonvolatile main memories in the rack.
We are also seeing the first tries at new processor architectures about to enter the datacenter: ARM 64 for cool/cold storage and web tier and OpenPower P8 for high power processing – multithreaded, multi-issue, pooled memory processing monsters. This is exciting to watch. There is also an emerging interest in application acceleration: general-purposing computing on graphics processing units (GPGPUs), regular expression processors (regex) live stream analytics, etc. We are also seeing the first generation of graph analysis deployed at massive scale in real time.
Innovation: The pace of innovation appears to be accelerating, although maybe I’m just getting older. But the easy gains are done. On one hand, datacenters need exponentially more compute and storage, and they need to operate 10x to 1000x more quickly. On the other, memory, processor cores, disks and flash technologies are getting no faster. The only way to fill that gap is through innovation. So it’s no surprise there are lots of interesting things happening at OEMs and ISVs, chip and solution companies, as well as open source community and startups. This is what makes it such an interesting time and industry.
Consumption shifts: We are seeing a decline in laptop and personal computer shipments, a drop that naturally is reducing storage demand in those markets. Laptops are also seeing a shift to SSD from HDD. This has been good for LSI, as our footprint in laptop HDDs had been small, but our presence in laptop SSDs is very strong. Smart phones and tablets are driving more cloud content, traffic and reliance on cloud storage. We have seen a dramatic increase in large HDDs for cloud storage, a trend that seems to be picking up speed, and we believe the cloud HDD market will be very healthy and will see the emergence of new, cloud-specific HDDs that are radically different and specifically designed for cool and cold storage.
There is also an explosion of SSD and PCIe flash cards in cloud computing for databases, caches, low-latency access and virtual machine (VM) enablement. Many applications that we take for granted would not be possible without these extreme low-latency, high-capacity flash products. But very few companies can make a viable storage system from flash at an acceptable cost, opening up an opportunity for many startups to experiment with different solutions.
Summary: So I believe the biggest hyperscale innovations are autonomous behavior and orchestration, HA at the datacenter level vs. machine level, and big data. These are radically changing the whole industry. And what are those changes for our industry and ecosystem? You name it: timing, end customers, new players, open initiatives, new architectures and algorithms, innovation, and consumption patterns. All that’s staying the same are legacy products and solutions.
These were great questions. Sometimes you need the press or someone outside the industry to ask a question that makes you step back and think about what’s going on. Great questions.
Tags: Alibaba, Apple Siri, Arista, ARM 64, Baidu, big data, Casandra, CEO & CIO Magazine, China, cloud storage, Cloudera, cold storage, cool storage, datacenter, datacenter ecosystem, Dell, flash, Foxconn, Google File System, Google MapReduce, Hadoop, hard disk drive, HDD, high availability, HP, hyperscale datacenter, Intel, Internet, latency, Microsoft, Mirantis, Nebula, OEM, Open Compute, OpenPower P8, OpenStack, original equipment manufacturer, Quanta, rack scale, RAID, Redhat Linux, SAS, SDDC, SDN, SDS, Serial Attached SCSI, software-defined datacenter, software-defined networks, software-defined storage, solid state drive, SSD, Tencent, TYAN, United Stack, virtual machine, VM, Wistron
Last week at LSI’s annual Accelerating Innovation Summit (AIS) the company took the wraps off a vision that should lead its technical direction for the next few years.
The LSI keynote featured a video of three situations as they might evolve in the future:
I’ll focus on just one of these to show how LSI expects the future to develop. In the bicycle accident scenario, a businessman falls to the ground while riding a bicycle in a foreign country. Security cameras that have been upgraded to understand what they see notify an emergency services agency which sends an ambulance to the scene. The paramedic performs a retinal scan on the victim, using it to retrieve his medical records, including his DNA sequence, from the web.
The businessman’s wearable body monitoring system also communicates with the paramedic’s instruments to share his vital signs. All of this information is used by cloud-based computers to determine a course of action which, in the video, requires an injection that has been custom-tuned to the victim’s current situation, his medical history, and his genetic makeup.
That’s a pretty tall order, and it will require several advances in the state of the art, but LSI is using this and other scenarios to work with its clients and translate this vision into the products of the future.
What are the key requirements to make this happen? Talwalkar told the audience that we need to create a society that is supported by preventive, predictive and assisted analytics to move in a direction where the general welfare is assisted by all that the Internet and advanced computing have to offer. Since data is growing at an exponential rate, he argued that this will require the instant retrieval of interlinked data objects at scale. Everything that is key to solving the task must be immediately available, and must be quickly analyzed to provide a solution to the problem at hand. The key will be the ability to process interlinked pieces of data that have not been previously structured to handle any particular situation.
To achieve this we will need larger-scale computing resources than are currently available, all closely interconnected, that all operate at very high speeds. LSI hopes to tap into these needs through its strengths in networking and communications chips for the communications, its HDD and server and storage connectivity array chips and boards for large-scale data, and its flash controller memory and PCIe SSD expertise for high performance.
LSI brought to AIS several of the customers and partners it is working with using to develop these technologies. Speakers from Intel, Microsoft, IBM, Toshiba, Ericsson and others showed how they are working with LSI’s various technologies to improve the performance of their own systems. On the exhibition floor booths from LSI and many of its clients demonstrated new technologies that performed everything from high-speed stock market analysis to fast flash management.
It’s pretty exciting to see a company that has a clear vision of its future and is committed to moving its entire ecosystem ahead to make that happen and help companies manage their business more effectively during what LSI calls the “Datacentric Era.” LSI has certainly put a lot of effort into creating a vision and determining where its talents can be brought to bear to improve our lives in the future.
Tags: AIS, chips, communications, connectivity, data, Datacentric Era, Ericsson, flash, flash memory, hard disk drive, HDD, IBM, Intel, large-scale data, Microsoft, Networking, server, Storage, Toshiba
The lifeblood of any online retailer is the speed of its IT infrastructure. Shoppers aren’t infinitely patient. Sluggish infrastructure performance can make shoppers wait precious seconds longer than they can stand, sending them fleeing to other sites for a faster purchase. Our federal government’s halting rollout of the Health Insurance Marketplace website is a glaring example of what can happen when IT infrastructure isn’t solid. A few bad user experiences that go viral can be damaging enough. Tens of thousands can be crippling.
In hyperscale datacenters, any number of problems including network issues, insufficient scaling and inconsistent management can undermine end users’ experience. But one that hits home for me is the impact of slow storage on the performance of databases, where the data sits. With the database at the heart of all those online transactions, retailers can ill afford to have their tier of database servers operating at anything less than peak performance.
Slow storage undermines database performance
Typically, Web 2.0 and e-commerce companies run relational databases (RDBs) on these massive server-centric infrastructures. (Take a look at my blog last week to get a feel for the size of these hyperscale datacenter infrastructures). If you are running that many servers to support millions of users, you are likely using some kind of open-sourced RDB such as MySQL or other variations. Keep in mind that Oracle 11gR2 likely retails around $30K per core but MSQL is free. But the performance of both, and most other relational databases, suffer immensely when transactions are retrieving data from storage (or disk). You can only throw so much RAM and CPU power at the performance problem … sooner rather than later you have to deal with slow storage.
Almost everyone in industry – Web 2.0, cloud, hyperscale and other providers of massive database infrastructures – is lining up to solve this problem the best way they can. How? By deploying flash as the sole storage for database servers and applications. But is low-latency flash enough? For sheer performance it beats rotational disk hands down. But … even flash storage has its limitations, most notably when you are trying to drive ultra-low latencies for write IOs. Most IO accesses by RDBs, which do the transactional processing, are a mix or read/writes to the storage. Specifically, the mix is 70%/30% reads/writes. These are also typically low q-depth accesses (less than 4). It is those writes that can really slow things down.
PCIe flash reduces write latencies
The good news is that the right PCIe flash technology in the mix can solve the slowdowns. Some interesting PCIe flash technologies designed to tackle this latency problem are on display at AIS this week. DRAM and in particular NVDRAM are being deployed as a tier in front of flash to really tackle those nasty write latencies.
Among other demos, we’re showing how a Nytro™ 6000 series PCIe flash card helps solve the MySQL database performance issues. The typical response time for a small data read (this is what the database will see for a Database IO) from an HDD is 5ms. Flash-based devices such as the Nytro WarpDrive® card can complete the same read in less than 50μs on average during testing, an improvement of several orders-of-magnitude in response time. This response time translates to getting much higher transactions out of the same infrastructure – but with less space (flash is denser) and a lot less power (flash consumes a lot lower power than HDDs).
We’re also showing the Nytro 7000 series PCIe flash cards. They reach even lower write latencies than the 6000 series and very low q-depths. The 7000 series cards also provide DRAM buffering while maintaining data-integrity even in the event of a power loss.
For online retailers and other businesses, higher database speeds mean more than just faster transactions. They can help keep those cash registers ringing.
Tags: AIS, database, DRAM, e-commerce, flash, flash memory, hard disk drive, HDD, hyperscale datacenter, latency, MySQL, NVDRAM, Nytro 6000, Nytro 7000, Nytro WarpDrive, Oracle, PCIe flash, relational database, storage latency, web 2.0, write latency
I often think about green, environmental impact, and what we’re doing to the environment. One major reason I became an engineer was to leave the world a little better than when I arrived. I’ve gotten sidetracked a few times, but I’ve tried to help, even if just a little.
The good people in LSI’s EHS (Environment, Health & Safety) asked me a question the other day about carbon footprint, energy impact, and materials use. Which got me thinking … OK – I know most people in LSI don’t really think of ourselves as a “green tech” company. But we are – really. No foolin’. We are having a big impact on the global power consumption and material consumption of the IT industry. And I mean that in a good way.
There are many ways to look at this, both from what we enable datacenters to do, to what we enable integrators to do, all the way to hard-core technology improvements and massive changes in what it’s possible to do.
Back in 2008 I got to speak at the AlwaysOn GoingGreen conference. (I was lucky enough to be just after Elon Musk– he’s a lot more famous now with Tesla doing so well.
http://www.smartplanet.com/video/making-the-case-for-green-it/305467 (at 2:09 in video)
IT consumes massive amounts of energy
The massive deployment of IT equipment, all the ancillary metal, plastic wiring, etc. that goes with them, consumes energy as its being shipped and moved halfway around the world, and, more importantly, then gets scrapped out quickly. This has been a concern for me for quite a while. I mean – think about that. As an industry we are generating about 9 million servers a year, about 3 million go into hyperscale datacenters (or hyperscale if you prefer). Many of those are scrapped on a 2, 3 or 4 year cycle – so in steady state, maybe 1 million to 2 million a year are scrapped. Worse – there is amazing use of energy by that many servers (even as they have advanced the state of the art unbelievably since 2008). And frankly, you and I are responsible for using all that power. Did you know thousands of servers are activated every time you make a Google® query from your phone?
I want to take a look at basic silicon improvements we make, the impact of disk architecture improvement, SSDs, system and improvements, efficiency improvements, and also where we’re going in the near future with eliminating scrap in hard drives and batteries. In reality, it’s the massive pressure on work/$ that has made us optimize everything – being able to do much more work at a lower cost, when a lot of cost is the energy and material that goes into the products that forces our hand. But the result is a real, profound impact on our carbon footprint that we should be proud of.
Sure we have a general silicon roadmap where each node enables reduced power, even as some standards and improvements actually increase individual device power. For example, our transition from 28nm semi process to 14nm FinFET can literally cut the power consumption of a chip in half. But that’s small potatoes.
How about Ethernet? It’s everywhere – right? Did you know servers often have 4 ethernet ports, and that there are a matching 4 ports on a network switch? LSI pioneered something called Energy Efficient Ethernet (EEE). We’re also one of the biggest manufacturers of Ethernet PHYs – the part that drives the cable – and we come standard in everything from personal computers to servers to enterprise switches. The savings are hard to estimate, because they depend very much on how much traffic there is, but you can realistically save Watts per interface link, and there are often 256 links in a rack. 500 Watts per rack is no joke, and in some datacenters it adds up to 1 or 2 MegaWatts.
How about something a little bigger and more specific? Hard disk drives. Did you know a typical hyperscale datacenter has between 1 million and 1.5 million disk drives? Each one of those consumes about 9 Watts, and most have 2 TBytes of capacity. So for easy math, 1 million drives is about 9 MegaWatts (!?) and about 2 Exabytes of capacity (remember – data is often replicated 3 or more times). Data capacities in these facilities are needed to grow about 50% per year. So if we did nothing, we would need to go from 1 million drives to 1.5 million drives: 9 MegaWatts goes to 13.5 MegaWatts. Wow! Instead – our high linearity, low noise PA and read channel designs are allowing drives to go to 4 TBytes per drives. (Sure the chip itself may use slightly more power, but that’s not the point, what it enables is a profound difference.) So to get that 50% increase in capacity we could actually reduce the number of drives deployed, with a net savings of 6.75 MegaWatts. Consider an average US home, with air conditioning, uses 1 kiloWatt. That’s almost 7,000 homes. In reality – they won’t get deployed that way – but it will still be a huge savings. Instead of buying another 0.5 million drives they would buy 0.25 million drives with a net savings of 2.2 MegaWatts. That’s still HUGE! (way to go, guys!) How many datacenters are doing that? Dozens. So that’s easily 20 or 30 MegaWatts globally. Did I say we saved them money too? A lot of money.
SSDs sip power to help improve energy profile
SSDs don’t always get the credit they deserve. Yes, they really are fast, and they are awesome in your laptop, but they also end up being much lower power than hard drives. Our controllers were in about half the flash solutions shipped last year. Think tens of millions. If you just assume they were all laptop SSDs (at least half were not) then that’s another 20 MegaWatts in savings.
Did you know that in a traditional datacenter, about 30% of the power going into the building is used for air conditioning? It doesn’t actually get used on the IT equipment at all, but is used to remove the heat that the IT equipment generates. We design our solutions so they can accommodate 40C ambient inlet air (that’s a little over 100F… hot). What that means is that the 30% of power used for the air conditioners disappears. Gone. That’s not theoretical either. Most of the large social media, search engine, web shopping, and web portal companies are using our solutions this way. That’s a 30% reduction in the power of storage solutions globally. Again, its MegaWatts in savings. And mega money savings too.
But let’s really get to the big hitters: improved work per server. Yep – we do that. In fact adding a Nytro™ MegaRAID® solution will almost always give you 4x the work out of a server. It’s a slam dunk if you’re running a database. You heard me – 1 server doing the work that it previously took 4 servers to do. Not only is that a huge savings in dollars (especially if you pay for software licenses!) but it’s a massive savings in power. You can replace 4 servers with 1, saving at least 900 Watts, and that lone server that’s left is actually dissipating less power too, because it’s actively using fewer HDDs, and using flash for most traffic instead. If you go a step further and use Nytro WarpDrive Flash cards in the servers, you can get much more – 6 to 8 times the work. (Yes, sometimes up to 10x, but let’s not get too excited). If you think that’s just theoretical again, check your Facebook® account, or download something from iTunes®. Those two services are the biggest users of PCIe® flash in the world. Why? It works cost effectively. And in case you haven’t noticed those two companies like to make money, not spend it. So again, we’re talking about MegaWatts of savings. Arguably on the order of 150 MegaWatts. Yea – that’s pretty theoretical, because they couldn’t really do the same work otherwise, but still, if you had to do the work in a traditional way, it would be around that.
It’s hard to be more precise than giving round numbers at these massive scales, but the numbers are definitely in the right zone. I can say with a straight face we save the world 10’s, and maybe even 100’s of MegaWatts per year. But no one sees that, and not many people even think about it. Still – I’d say LSI is a green hero.
Hey – we’re not done by a long shot. Let’s just look at scrap. If you read my earlier post on false disk failure, you’ll see some scary numbers. (http://blog.lsi.com/what-is-false-disk-failure-and-why-is-it-a-problem/ ) A normal hyperscale datacenter can expect 40-60 disks per day to be mistakenly scrapped out. That’s around 20,000 disk drives a year that should not have been scrapped, from just one web company. Think of the material waste, shipping waste, manufacturing waste, and eWaste issues. Wow – all for nothing. We’re working on solutions to that. And batteries. Ugly, eWaste, recycle only, heavy metal batteries. They are necessary for RAID protected storage systems. And much of the world’s data is protected that way – the battery is needed to save meta-data and transient writes in the event of a power failure, or server failure. We ship millions a year. (Sorry, mother earth). But we’re working diligently to make that a thing of the past. And that will also result in big savings for datacenters in both materials and recycling costs.
Can we do more? Sure. I know I am trying to get us the core technologies that will help reduce power consumption, raise capability and performance, and reduce waste. But we’ll never be done with that march of technology. (Which is a good thing if engineering is your career…)
I still often think about green, environmental impact, and what we’re doing to the environment. And I guess in my own small way, I am leaving the world a little better than when I arrived. And I think we at LSI should at least take a moment and pat ourselves on the back for that. You have to celebrate the small victories, you know? Even as the fight goes on.
I want to warn you, there is some thick background information here first. But don’t worry. I’ll get to the meat of the topic and that’s this: Ultimately, I think that PCIe® cards will evolve to more external, rack-level, pooled flash solutions, without sacrificing all their great attributes today. This is just my opinion, but other leaders in flash are going down this path too…
I’ve been working on enterprise flash storage since 2007 – mulling over how to make it work. Endurance, capacity, cost, performance have all been concerns that have been grappled with. Of course the flash is changing too as the nodes change: 60nm, 50nm, 35nm, 24nm, 20nm… and single level cell (SLC) to multi level cell (MLC) to triple level cell (TLC) and all the variants of these “trimmed” for specific use cases. The spec “endurance” has gone from 1 million program/erase cycles (PE) to 3,000, and in some cases 500.
It’s worth pointing out that almost all the “magic” that has been developed around flash was already scoped out in 2007. It just takes a while for a whole new industry to mature. Individual die capacity increased, meaning fewer die are needed for a solution – and that means less parallel bandwidth for data transfer… And the “requirement” for state-of-the-art single operation write latency has fallen well below the write latency of the flash itself. (What the ?? Yea – talk about that later in some other blog. But flash is ~1500uS write latency, where state of the art flash cards are ~50uS.) When I describe the state of technology it sounds pretty pessimistic. I’m not. We’ve overcome a lot.
We built our first PCIe card solution at LSI in 2009. It wasn’t perfect, but it was better than anything else out there in many ways. We’ve learned a lot in the years since – both from making them, and from dealing with customer and users – about our own solutions and our competitors. We’re lucky to be an important player in storage, so in general the big OEMs, large enterprises and the hyperscale datacenters all want to talk with us – not just about what we have or can sell, but what we could have and what we could do. They’re generous enough to share what works and what doesn’t. What the values of solutions are and what the pitfalls are too. Honestly? It’s the hyperscale datacenters in the lead both practically and in vision.
If you haven’t nodded off to sleep yet, that’s a long-winded way of saying – things have changed fast, and, boy, we’ve learned a lot in just a few years.
Most important thing we’ve learned…
Most importantly, we’ve learned it’s latency that matters. No one is pushing the IOPs limits of flash, and no one is pushing the bandwidth limits of flash. But they sure are pushing the latency limits.
PCIe cards are great, but…
We’ve gotten lots of feedback, and one of the biggest things we’ve learned is – PCIe flash cards are awesome. They radically change performance profiles of most applications, especially databases allowing servers to run efficiently and actual work done by that server to multiply 4x to 10x (and in a few extreme cases 100x). So the feedback we get from large users is “PCIe cards are fantastic. We’re so thankful they came along. But…” There’s always a “but,” right??
It tends to be a pretty long list of frustrations, and they differ depending on the type of datacenter using them. We’re not the only ones hearing it. To be clear, none of these are stopping people from deploying PCIe flash… the attraction is just too compelling. But the problems are real, and they have real implications, and the market is asking for real solutions.
Of course, everyone wants these fixed without affecting single operation latency, or increasing cost, etc. That’s what we’re here for though – right? Solve the impossible?
A quick summary is in order. It’s not looking good. For a given solution, flash is getting less reliable, there is less bandwidth available at capacity because there are fewer die, we’re driving latency way below the actual write latency of flash, and we’re not satisfied with the best solutions we have for all the reasons above.
If you think these through enough, you start to consider one basic path. It also turns out we’re not the only ones realizing this. Where will PCIe flash solutions evolve over the next 2, 3, 4 years? The basic goals are:
One easy answer would be – that’s a flash SAN or NAS. But that’s not the answer. Not many customers want a flash SAN or NAS – not for their new infrastructure, but more importantly, all the data is at the wrong end of the straw. The poor server is left sucking hard. Remember – this is flash, and people use flash for latency. Today these SAN type of flash devices have 4x-10x worse latency than PCIe cards. Ouch. You have to suck the data through a relatively low bandwidth interconnect, after passing through both the storage and network stacks. And there is interaction between the I/O threads of various servers and applications – you have to wait in line for that resource. It’s true there is a lot of startup energy in this space. It seems to make sense if you’re a startup, because SAN/NAS is what people use today, and there’s lots of money spent in that market today. However, it’s not what the market is asking for.
Another easy answer is NVMe SSDs. Right? Everyone wants them – right? Well, OEMs at least. Front bay PCIe SSDs (HDD form factor or NVMe – lots of names) that crowd out your disk drive bays. But they don’t fix the problems. The extra mechanicals and form factor are more expensive, and just make replacing the cards every 5 years a few minutes faster. Wow. With NVME SSDs, you can fit fewer HDDs – not good. They also provide uniformly bad cooling, and hard limit power to 9W or 25W per device. But to protect the storage in these devices, you need to have enough of them that you can RAID or otherwise protect. Once you have enough of those for protection, they give you awesome capacity, IOPs and bandwidth, too much in fact, but that’s not what applications need – they need low latency for the working set of data.
What do I think the PCIe replacement solutions in the near future will look like? You need to pool the flash across servers (to optimize bandwidth and resource usage, and allocate appropriate capacity). You need to protect against failures/errors and limit the span of failure, commit writes at very low latency (lower than native flash) and maintain low latency, bottleneck-free physical links to each server… To me that implies:
That means the performance looks exactly as if each server had multiple PCIe cards. But the capacity and bandwidth resources are shared, and systems can remain resilient. So ultimately, I think that PCIe cards will evolve to more external, rack level, pooled flash solutions, without sacrificing all their great attributes today. This is just my opinion, but as I say – other leaders in flash are going down this path too…
What’s your opinion?
Tags: DAS, datacenter, direct attached storage, enterprise IT, flash, hard disk drive, HDD, hyperscale, latency, NAS, network attached storage, NVMe, PCIe, SAN, solid state drive, SSD, storage area network
Big data and Hadoop are all about exploiting new value and opportunities with data. In financial trading, business and some areas of science, it’s all about being fastest or first to take advantage of the data. The bigger the data sets, the smarter the analytics. The next competitive edge with big data comes when you layer in flash acceleration. The challenge is scaling performance in Hadoop clusters.
The most cost-effective option emerging for breaking through disk-to-I/O bottlenecks to scale performance is to use high-performance read/write flash cache acceleration cards for caching. This is essentially a way to get more work for less cost, by bringing data closer to the processing. The LSI® Nytro™ product has been shown during testing to improve the time it takes to complete Hadoop software framework jobs up to a 33%.
Flash cache cards increase Hadoop application performance
Combining flash cache acceleration cards with Hadoop software is a big opportunity for end users and suppliers. LSI estimates that less than 10% of Hadoop software installations today incorporate flash acceleration1. This will grow rapidly as companies see the increased productivity and ROI of flash to accelerate their systems. And Hadoop software adoption is also growing fast. IDC predicts a CAGR of as much as 60% by 20162. Drivers include IT security, e-commerce, fraud detection and mobile data user management. Gartner predicts that Hadoop software will be in two-thirds of advanced analytics products by 20153. Many thousands of Hadoop software clusters are already deployed.
Where flash makes the most immediate sense is with those who have smaller clusters doing lots of in-place batch processing. Hadoop is purpose-built for analyzing a variety of data, whether structured, semi-structured or unstructured, without the need to define a schema or otherwise anticipate results in advance. Hadoop enables scaling that allows an unprecedented volume of data to be analyzed quickly and cost-effectively on clusters of commodity servers. Speed gains are about data proximity. This is why flash cache acceleration typically delivers the highest performance gains when the card is placed directly in the server on the PCI Express® (PCIe) bus.
Combining the best of flash and HDDs to drive higher performance and storage capacity
PCIe flash cache cards are now available with multiple terabytes of NAND flash storage, which substantially increases the hit rate. We offer a solution with both onboard flash modules and Serial-Attached SCSI (SAS) interfaces to enable high-performance direct-attached storage (DAS) configurations consisting of solid state and hard disk drive storage. This couples the low-latency performance benefits of flash with the capacity and cost-per-gigabyte advantages of HDDs.
To keep the processor close to the data, Hadoop uses servers with DAS. And to get the data even closer to the processor, the servers are usually equipped with significant amounts of random access memory (RAM). An additional benefit: Smart implementation of Hadoop and flash components can reduce the overall server footprint and simplify scaling, with some solutions enabling up to 128 devices to share a very high bandwidth interface. Most commodity servers provide 8 or less SATA ports for disks, reducing expandability.
Hadoop is great, but flash-accelerated Hadoop is best. It’s an effective way, as you work to extract full value from big data, to secure a competitive edge.
It may sound crazy, but hard disk drives (HDDs) do not have a delete command. Now we all know HDDs have a fixed capacity, so over time the older data must somehow get removed, right? Actually it is not removed, but overwritten. The operating system (OS) uses a reference table to track the locations (addresses) of all data on the HDD. This table tells the OS which spots on the HDD are used and which are free. When the OS or a user deletes a file from the system, the OS simply marks the corresponding spot in the table as free, making it available to store new data.
The HDD is told nothing about this change, and it does not need to know since it would not do anything with that information. When the OS is ready to store new data in that location, it just sends the data to the HDD and tells it to write to that spot, directly overwriting the prior data. It is simple and efficient, and no delete command is required.
However, with the advent of NAND flash-based solid state drives (SSDs) a new problem emerged. In my blog, Gassing up your SSD, I explain how NAND flash memory pages cannot be directly overwritten with new data, but must first be erased at the block level through a process called garbage collection (GC). I further describe how the SSD uses non-user space in the flash memory (over provisioning or OP) to improve performance and longevity of the SSD. In addition, any user space not consumed by the user becomes what we call dynamic over provisioning – dynamic because it changes as the amount of stored data changes.
When less data is stored by the user, the amount of dynamic OP increases, further improving performance and endurance. The problem I alluded to earlier is caused by the lack of a delete command. Without a delete command, every SSD will eventually fill up with data, both valid and invalid, eliminating any dynamic OP. The result would be the lowest possible performance at that factory OP level. So unlike HDDs, SSDs need to know what data is invalid in order to provide optimum performance and endurance.
Keeping your SSD TRIM
A number of years ago, the storage industry got together and developed a solution between the OS and the SSD by creating a new SATA command called TRIM. It is not a command that forces the SSD to immediately erase data like some people believe. Actually the TRIM command can be thought of as a message from the OS about what previously used addresses on the SSD are no longer holding valid data. The SSD takes those addresses and updates its own internal map of its flash memory to mark those locations as invalid. With this information, the SSD no longer moves that invalid data during the GC process, eliminating wasted time rewriting invalid data to new flash pages. It also reduces the number of write cycles on the flash, increasing the SSD’s endurance. Another benefit of the TRIM command is that more space is available for dynamic OP.
Today, most current operating systems and SSDs support TRIM, and all SandForce Driven™ member SSDs have always supported TRIM. Note that most RAID environments do not support TRIM, although some RAID 0 configurations have claimed to support it. I have presented on this topic in detail previously. You can view the presentation in full here. In my next blog I will explain how there may be an alternate solution using SandForce Driven member SSDs.
I’ve been travelling to China quite a bit over the last year or so. I’m sitting in Shenzhen right now (If you know Chinese internet companies, you’ll know who I’m visiting). The growth is staggering. I’ve had a bit of a trains, planes, automobiles experience this trip, and that’s exposed me to parts of China I never would have seen otherwise. Just to accommodate sheer population growth and the modest increase in wealth, there is construction everywhere – a press of people and energy, constant traffic jams, unending urban centers, and most everything is new. Very new. It must be exciting to be part of that explosive growth. What a market. I mean – come on – there are 1.3 billion potential users in China.
The amazing thing for me is the rapid growth of hyperscale datacenters in China, which is truly exponential. Their infrastructure growth has been 200%-300% CAGR for the past few years. It’s also fantastic walking into a building in China, say Baidu, and feeling very much at home – just like you walked into Facebook or Google. It’s the same young vibe, energy, and ambition to change how the world does things. And it’s also the same pleasure – talking to architects who are super-sharp, have few technical prejudices, and have very little vanity – just a will to get to business and solve problems. Polite, but blunt. We’re lucky that they recognize LSI as a leader, and are willing to spend time to listen to our ideas, and to give us theirs.
Even their infrastructure has a similar feel to the US hyperscale datacenters. The same only different. ;-)
A lot of these guys are growing revenue at 50% per year, several getting 50% gross margin. Those are nice numbers in any country. One has $100’s of billions in revenue. And they’re starting to push out of China. So far their pushes into Japan have not gone well, but other countries should be better. They all have unique business models. “We” in the US like to say things like “Alibaba is the Chinese eBay” or “Sina Weibo is the Chinese Twitter”…. But that’s not true – they all have more hybrid business models, unique, and so their datacenter goals, revenue and growth have a slightly different profile. And there are some very cool services that simply are not available elsewhere. (You listening Apple®, Google®, Twitter®, Facebook®?) But they are all expanding their services, products and user base. Interestingly, there is very little public cloud in China. So there are no real equivalents to Amazon’s services or Microsoft’s Azure. I have heard about current development of that kind of model with the government as initial customer. We’ll see how that goes.
100’s of thousands of servers. They’re not the scale of Google, but they sure are the scale of Facebook, Amazon, Microsoft…. It’s a serious market for an outfit like LSI. Really it’s a very similar scale now to the US market. Close to 1 million servers installed among the main 4 players, and exabytes of data (we’ve blown past mere petabytes). Interestingly, they still use many co-location facilities, but that will change. More important – they’re all planning to probably double their infrastructure in the next 1-2 years – they have to – their growth rates are crazy.
Often 5 or 6 distinct platforms, just like the US hyperscale datacenters. Database platforms, storage platforms, analytics platforms, archival platforms, web server platforms…. But they tend to be a little more like a rack of traditional servers that enterprise buys with integrated disk bays, still a lot of 1G Ethernet, and they are still mostly from established OEMs. In fact I just ran into one OEM’s American GM, who I happen to know, in Tencent’s offices today. The typical servers have 12 HDDs in drive bays, though they are starting to look at SSDs as part of the storage platform. They do use PCIe® flash cards in some platforms, but the performance requirements are not as extreme as you might imagine. Reasonably low latency and consistent latency are the premium they are looking for from these flash cards – not maximum IOPs or bandwidth – very similar to their American counterparts. I think hyperscale datacenters are sophisticated in understanding what they need from flash, and not requiring more than that. Enterprise could learn a thing or two.
Some server platforms have RAIDed HDDs, but most are direct map drives using a high availability (HA) layer across the server center – Hadoop® HDFS or self-developed Hadoop like platforms. Some have also started to deploy microserver archival “bit buckets.” A small ARM® SoC with 4 HDDs totaling 12 TBytes of storage, giving densities like 72 TBytes of file storage in 2U of rack. While I can only find about 5,000 of those in China that are the first generation experiments, it’s the first of a growing wave of archival solutions based on lower performance ARM servers. The feedback is clear – they’re not perfect yet, but the writing is on the wall. (If you’re wondering about the math, that’s 5,000 x 12 TBytes = 60 Petabytes….)
Yes, it’s important, but maybe more than we’re used to. It’s harder to get licenses for power in China. So it’s really important to stay within the envelope of power your datacenter has. You simply can’t get more. That means they have to deploy solutions that do more in the same power profile, especially as they move out of co-located datacenters into private ones. Annually, 50% more users supported, more storage capacity, more performance, more services, all in the same power. That’s not so easy. I would expect solar power in their future, just as Apple has done.
Here’s where it gets interesting. They are developing a cousin to OpenCompute that’s called Scorpio. It’s Tencent, Alibaba, Baidu, and China Telecom so far driving the standard. The goals are similar to OpenCompute, but more aligned to standardized sub-systems that can be co-mingled from multiple vendors. There is some harmonization and coordination between OpenCompute and Scorpio, and in fact the Scorpio companies are members of OpenCompute. But where OpenCompute is trying to change the complete architecture of scale-out clusters, Scorpio is much more pragmatic – some would say less ambitious. They’ve finished version 1 and rolled out about 200 racks as a “test case” to learn from. Baidu was the guinea pig. That’s around 6,000 servers. They weren’t expecting more from version 1. They’re trying to learn. They’ve made mistakes, learned a lot, and are working on version 2.
Even if it’s not exciting, it will have an impact because of the sheer size of deployments these guys are getting ready to roll out in the next few years. They see the progression as 1) they were using standard equipment, 2) they’re experimenting and learning from trial runs of Scorpio versions 1 and 2, and then they’ll work on 3) new architectures that are efficient and powerful, and different.
Information is pretty sketchy if you are not one of the member companies or one of their direct vendors. We were just invited to join Scorpio by one of the founders, and would be the first group outside of China to do so. If that all works out, I’ll have a much better idea of the details, and hopefully can influence the standards to be better for these hyperscale datacenter applications. Between OpenCompute and Scorpio we’ll be seeing a major shift in the industry – a shift that will undoubtedly be disturbing to a lot of current players. It makes me nervous, even though I’m excited about it. One thing is sure – just as the server market volume is migrating from traditional enterprise to hyperscale datacenter (25-30% of the server market and growing quickly), we’re starting to see a migration to Chinese hyperscale datacenters from US-based ones. They have to grow just to stay still. I mean – come on – there are 1.3 billion potential users in China….
Tags: Alibaba, Amazon, Apple, ARM, Baidu, China, China Telecom, datacenter, Facebook, Google, Hadoop, hard disk drive, HDD, hyperscale, Microsoft, OpenCompute, Scorpio, Shenzhen, Sina Weibo, solid state drive, SSD, Tencent, Twitter
There’s no need to wait for higher speed. Server builders can take advantage of 12Gb/s SAS now. And this is even as HDD and SSD makers continue to tweak, tune and otherwise prepare their 12Gb/s SAS products for market. The next generation of 12Gb/s SAS without supporting drives? What gives?
It’s simple. LSI is already producing 12Gb/s ROC and IOC solutions, meaning that customers can take advantage of 12Gb/s SAS performance today with currently shipping systems and storage. As for the numbers, LSI 12Gb/s SAS enables performance increases of up to 45% in throughput and up to 58% in IOPS when compared to 6Gb/s SAS.
True, 12Gb/s SAS isn’t a Big Bang Disruption in storage systems; rather it’s an evolutionary change, but a big step forward. It may not be clear why it matters so much, so I want to briefly explain. In latest generation PCIe 3 systems, 6Gb/s SAS is the bottleneck that prevents systems from achieving full PCIe 3 throughput of 6,400 MB/s.
With 12Gb/s SAS, customers will be able to take full advantage of the performance of PCIe 3 systems. Earlier this month at CeBIT computer expo in Hanover, Germany, we announced that we are the first to ship production-level 12Gb/s SAS ROC (RAID on Chip) and IOC (I/O Controllers) to OEM customers. This convergence of new technologies and the expansion of existing capabilities create significant improvements for datacenters of all kinds.
At CeBIT, we demonstrated our 12Gb/s SAS solutions with the unique DataBoltTM feature and how, with DataBolt, systems with 6Gb/s SAS HDDs can achieve 12Gb/s SAS performance.
DataBolt uses bandwidth aggregation to create throughput performance acceleration. Most importantly, customers don’t have to wait for the next inflection in drive design to get the highest possible performance and connectivity.
I’ve spent a lot of time with hyperscale datacenters around the world trying to understand their problems – and I really don’t care what area those problems are as long as they’re important to the datacenter. What is the #1 Real Problem for many hyperscale datacenters? It’s something you’ve probably never heard about, and probably have not even thought about. It’s called false disk failure. Some hyperscale datacenters have crafted their own solutions – but most have not.
Why is this important, you ask? Many large datacenters today have 1 million to 4 million hard disk drives (HDDs) in active operation. In anyone’s book that’s a lot. It’s also a very interesting statistical sample size of HDDs. Hyperscale datacenters get great pricing on HDDs. Probably better than OEMs get, and certainly better than the $79 for buying 1 HDD at your local Fry’s store. So you would imagine if a disk fails – no one cares – they’re cheap and easy to replace. But the burden of a failed disk is much more than the raw cost of the disk:
Let’s put some scale to this problem, and you’ll begin to understand the issue. One modest size hyperscale datacenter has been very generous in sharing its real numbers. (When I say modest, they are ~1/4 to 1/2 the size of many other hyperscale datacenters, but they are still huge – more than 200k servers). Other hyperscale datacenters I have checked with say – yep, that’s about right. And one engineer I know at an HDD manufacturer said – “wow – I expected worse than that. That’s pretty good.” To be clear – these are very good HDDs they are using, it’s just that the numbers add up.
The raw data:
RAIDed SAS HDDs
Non-RAIDed (direct map) SATA drives behind HBAs
What’s interesting is the relative failure rate of SAS drives vs. SATA. It’s about an order of magnitude worse in SATA drives than SAS. Frankly some of this is due to protocol differences. SAS allows far more error recovery capabilities, and because they also tend to be more expensive, I believe manufacturers invest in slightly higher quality electronics and components. I know the electronics we ship into SAS drives is certainly more sophisticated than SATA drives.
False fail? What? Yea, that’s an interesting topic. It turns out that about 40% of the time with SAS and about 50% of the time with SATA, the drive didn’t actually fail. It just lost its marbles for a while. When they pull the drive out and put it into a test jig, everything is just fine. And more interesting, when they put the drive back into service, it is no more statistically likely to fail again than any other drive in the datacenter. Why? No one knows. I suspect though.
I used to work on engine controllers. That’s a very paranoid business. If something goes wrong and someone crashes, you have a lawsuit on your hands. If a controller needs a recall, that’s millions of units to replace, with a multi-hundred dollar module, and hundreds of dollars in labor for each one replaced. No one is willing to take that risk. So we designed very carefully to handle soft errors in memory and registers. We incorporated ECC like servers use, background code checksums and scrubbing, and all sorts of proprietary techniques, including watchdogs and super-fast self-resets that could get operational again in less than a full revolution of the engine. Why? – the events were statistically rare. The average controller might see 1 or 2 events in its lifetime, and a turn of the ignition would reset that state. But the events do happen, and so do recalls and lawsuits… HDD controllers don’t have these protections, which is reasonable. It would be an inappropriate cost burden for their price point.
You remember the Toyota Prius accelerator problems? I know that controller was not protected for soft errors. And the source of the problem remained a “mystery.” Maybe it just lost its marbles for a while? A false fail if you will. Just sayin’.
Back to HDDs. False fail is especially frustrating, because half the HDDs actually didn’t need to be replaced. All the operational costs were paid for no reason. The disk just needed a power cycle reset. (OK, that introduces all sorts of complex management by the RAID controller or application to manage that 10 second power reset cycle and application traffic created in that time – be we can handle that.)
Daily, this datacenter has to:
And 1/2 of that is for no reason at all.
First – why not rebuild the disk if it’s RAIDed? Usually hyperscale datacenters use clustered applications. A traditional RAID rebuild drives the server performance to ~50%, and for a 2TByte drive, under heavy application load (definition of a hyperscale datacenter) can truly take up to a week. 50% performance for a week? In a cluster that means the overall cluster is running ~50% performance. Say 200 nodes in a cluster – that means you just lost ~100 nodes of work – or 50% of cluster performance. It’s much simpler to just take the node offline with the failed drive, and get 99.5% cluster performance, and operationally redistribute the workload across multiple nodes (because you have replicated data elsewhere). But after rebuild, the node will have to be re-synced or re-imaged. There are ways to fix all this. We’ll talk about them on another day. Or you can simply run direct mapped storage, and unmounts the failed drive.
Next – Why replicate data over the network, and why is that a big deal? For geographic redundancy (say a natural disaster at one facility) and regional locality, hyperscale datacenters need multiple data copies. Often 3 copies so they can do double duty as high-availability copies, or in the case of some erasure coding, 2.2 to 2.5 copies (yea – weird math – how do you have 0.5 copy…). When you lose one copy, you are down to 2, possibly 1. You need to get back to a reliable number again. Fast. Customers are loyal because of your perfect data retention. So you need to replicate that data and re-distribute it across the datacenter on multiple servers. That’s network traffic, and possibly congestion, which affects other aspects of the operations of the datacenter. In this datacenter it’s about 50 hours of 10G Ethernet traffic every day.
To be fair, there is a new standard in SAS interfaces that will facilitate resetting a disk in-situ. And there is the start of discussion of the same around SATA – but that’s more problematic. Whatever the case, it will be a years before the ecosystem is in place to handle the problems this way.
What’s that mean to you?
Well. You can expect something like 1/100 of your drives to really fail this year. And you can expect another 1/100 of your drives to fail this year, but not actually be failed. You’ll still pay all the operational overhead of not actually having a failed drive – rebuilds, disk replacements, management interventions, scheduled downtime/maintenance time, and the OEM replacement price for that drive – what $600 or so ?… Depending on your size, that’s either a don’t care, or a big deal. There are ways to handle this, and they’re not expensive – much less than the disk carrier you already pay for to allow you to replace that drive – and it can be handled transparently – just a log entry without seeing any performance hiccups. You just need to convince your OEM to carry the solution.