It has been 2 months into my new role at DDN as a Solutions Architect. With many revolving doors around me, I have been trying to find the essence, the critical cog of the data infrastructure that supports the accelerated computing of the Nvidia GPU clusters. The more I read and engage, a pattern emerged. I found that cog in the supercharged data paths between the storage infrastructure systems and the GPU clusters. I will share more.
To set the context, let me start with a wonderful article I read in CIO.com back in July 2024. It was titled “Storage: The unsung hero of AI deployments“. It was music to my ears because as a long-time practitioner in the storage technology industry, it is time the storage industry gets its credit it deserves.
What is the data path?
To put it simply, a Data Path, from a storage context, is the communication route taken by the data bits between the compute system’s processing and program memory and the storage subsystem. The links and the established sessions can be within the system components such as the PCIe bus or external to the system through the shared networking infrastructure.
High speed accelerated data paths
In the world of accelerated computing such as AI and HPC, there are additional, more advanced technologies to create even faster delivery of the data bits. This is the accelerated data paths between the compute nodes and the storage subsystems. Following on, I share a few of these technologies that are lesser used in the enterprise storage segment.
I am strong believer of using the right tool to do the job right. I have said this before 2 years ago, in my blog “Stating the case for a Storage Appliance approach“. It was written when I was previously working for an open source storage company. And I am an advocate of the crafter versus assembler mindset, especially in the enterprise and high- performance storage technology segments.
I have joined DDN. Even with DDN that same mindset does not change a bit. I have been saying all along that the storage appliance model should always be the mindset for the businesses’ peace-of-mind.
My view of the storage appliance model began almost 25 years. I came into NAS systems world via Sun Microsystems®. Sun was famous for running NFS servers on general Sun Solaris servers. NFS services on Unix systems. Back then, I remember arguing with one of the Sun distributors about the tenets of running NFS over 100Mbit/sec Ethernet on Sun servers. I was drinking Sun’s Kool-Aid big time.
When I joined Network Appliance® (now NetApp®) in 2000, my worldview of putting software on general purpose servers changed. Network Appliance®, had one product family, the FAS700 (720, 740, 760) family. All NetApp® did was to serve NFS services in the beginning. They were the NAS filers and nothing else.
I was completed sold on the appliance way with NetApp®. Firstly, it was my very first time knowing such network storage services could be provisioned with an appliance concept. This was different from Sun. I was used to managing NFS exports on a Sun SPARCstation 20 to Unix clients in the network.
Secondly, my mindset began to shape that “you have to have the right tool to the job correctly and extremely well“. Well, the toaster toasts bread very well and nothing else. And the fridge (an analogy used by Dave Hitz, I think) does what it does very well too. That is what the appliance does. You definitely cannot grill a steak with a bread toaster, just like you can’t run an excellent, ultra-high performance storage services to serve the demanding AI and HPC applications on a general server platform. You have to have a storage appliance solution for High-Speed Storage.
That little Network Appliance® toaster award given out to exemplary employees stood vividly in my mind. The NetApp® tagline back then was “Fast, Simple, Reliable”. That solidifies my mindset for the high-speed storage in AI and HPC projects in present times.
DDN AI400X2 Turbo Appliance
Costs Benefits and Risks
I like to think about what the end users are thinking about. There are investments costs involved, and along with it, risks to the investments as well as their benefits. Let’s just simplify and lump them into Cost-Benefits-Risk analysis triangle. These variables come into play in the decision making of AI and HPC projects.
NAS (network attached storage) is obviously the file-level workhorse for shared resources in the network of any organization.SMB (server message block) for Windows environments and NFS (network file system) for Linux platforms are the 2 most prominent protocols that rule the NAS world. Of course we have SMB implementations in the form of Samba and others in non-Windows, Linux and NFS implementations in Windows as well.
As the versions of both network file sharing protocols iterated, present versions of SMB v3.x and NFS v4.x (NFS v3 on the supported Linux kernel version) on the client-side have evolved well. Both now have enhanced resiliency and performance improvement features. And there is an underlying similarity of both implementations. This blog looks at the client-side architectures of both.
One TCP connection
NAS is a client-server architecture. Over the network, NAS clients (SMB or NFS) access their corresponding NAS server(s) – SMB or NFS server(s) – through the TCP/IP network.
NAS client-server architecture (Credit: https://hypertecsp.com/en-CA/knowledge-base/nas-vs-san/)
One very important key starting point to note is the use of one TCP connection per NAS client to the NAS server relationship. For both SMB and NFS, there is just one TCP link between client and the server even if there are several SMB mapped shares or NFS mount points respectively on the clients.
For a long time, this one TCP connection is sufficient for the NAS traffic. But as the network file accesses grow, this connection becomes both a single point of failure as well as a performance bottleneck.
Deduplication in OpenZFS has been a bugbear for some years now. As data sets get larger, they have become even more difficult in using the present DeDuplication Table (DDT) method. Deduplication in OpenZFS is often derided as overwhelming and sluggish in performance.
Moreover, there is a common folklore passed on and on about allocating 5GB of RAM for every 1TB to dedupe in OpenZFS. I don’t know where this “sizing” came about. Probably derived from something Jeff Bonwickwrote back in the early days of ZFS. But there is some truth to this “rule of thumb”, commonly passed around in the TrueNAS® circles.
Nevertheless, given the exponential growth of data, and the advancement of processing power in modern day computer systems, the OpenZFS development community has decided to revamp the DDT method. Several prominent luminaries from iXsystems™, Klara Systems and the OpenZFS community have got together in mid-2023 to develop FDT or Fast Dedupe Table. And we got to see FDT announced to the world in the most recent OpenZFS Developer Summit in November 2023.
Fast Dedupe Table (FDT)
Fast Dedupe Table (FDT) is a log-based dedupe. In OpenZFS, all the write block I/Os that come into OpenZFS are coalesced into transaction groups (TXGs), hashed and checksummed, before they are committed to persistent media.
The new implementation in FDT is to put these incoming TXGs checksums and hashes into an append-only log structure in persistent storage, and also tracking the hashed changes in an AVL-tree residing in the memory. An AVL tree is a self-balancing binary search tree structure that is very efficient in searching, thus giving FDT the speed in initiating the deduplication lookups and updates.
OpenZFS Fast Dedupe Table (FDT) in a nutshell
The append-only log structure works hand-in-hand with the AVL tree to accept and stage (including intelligent sorting) the hash entries that are coming in after the TXGs writes. Then at a certain marker, that could be at a particular time-based trigger or a high-water mark, then the entries in the logs and AVL tree are flushed to the ZAP (ZFS Attribute Processor) where the actual full map of the OpenZFS blocks reside.
I was listening to several storage luminaries in the GestaltIT’s podcast “No one understands Storage anymore” a few of weeks ago. Around the minute of 11.09 in the podcast, Dr. J. Metz, SNIA® Chair, brought up this is powerful quote “Storage does not mean Capacity“. It struck me, not in a funny way. It is what it is, and it something I wanted to say to many who do not understand the storage solutions they are purchasing. It exemplifies what is wrong in the many organizations today in their understanding of investing in a storage infrastructure project.
This is my pet peeve. The first words uttered in most, if not all storage requirements in my line of work are, “I want this many Terabytes of storage“. There are no other details and context of what the other requirement factors are, such as availability, performance, future growth, etc. Or even the goals to achieve when purchasing a storage system and operating it. What is the improvement they are looking for?What are the problems to solve?
Where is the OKR?
It pains me to say this. For the folks who have in the IT industry for years, both end users and IT purveyors alike, most are absolutely clueless about OKR (Objectives and Key Results) for their storage infrastructure project. Many cannot frame the data challenges they are facing, and they have no idea where to go next. There is no alignment. There is no strategy. Even worse, there is no concept of how their storage infrastructure investments will improve their business and operations.
Just the other day, one company director from a renown IT integrator here in Malaysia came calling. He has been in the IT industry since 1989 (I checked his Linkedin profile), asking to for a 100TB storage quote. I asked a few questions about availability, performance, scalability; the usual questions a regular IT guy would ask. He has no idea, and instead of telling me he didn’t know, he gave me a runaround of this and that. Plenty of yada, yada nonsense.
In the end, I told him to buy a consumer grade storage appliance from Taiwan. I will just let him make a fool of himself in front of his customer since he didn’t want to take accountability of ensuring his customer get a proper enterprise storage solution in good faith. His customer is probably in the same mould as well.
Defensive Strategies as Data Foundations
A strong storage infrastructure foundation is vital for good Data Credibility. If you do the right things for your data, there is Data Value, and it will serve your business well. Both Data Credibility and Data Value create confidence. And Confidence equates Trust.
In order to create the defensive strategies let’s look at storage Availability, Protection, Accessibility, Management Security and Compliance. These are 6 of the 8 data points of the A.P.P.A.R.M.S.C. framework.
Offensive Strategies as Competitive Advantage
Once we have achieved stability of the storage infrastructure foundation, then we can turn over and drive towards storage Performance, Recovery, plus things like Scalability and Agility.
With a strong data infrastructure foundation, the organization can embark on the offensive, and begin their business transformation journey, knowing that their data is well run, protection, and performs.
Alignment with Data and Business Goals
Why are the defensive and offensive strategies requiring alignment to business goals?
The fact is simple. It is about improving the business and operations, and setting OKRs is key to measure the ROI (return of investment) of getting the storage systems and the solutions in place. It is about switching the cost-fearing (negative) mindset to a profit-conviction (positive) mindset.
For example, maybe the availability of the data to the business is poor. Maybe there is the need to have access to the data 24×7, because the business is going online. The simple measurable fact is we can move availability from 95% uptime to 99.99% uptime with an HA storage system.
Perhaps there are concerns about recoverability in the deluge of ransomware threats. Setting new RPO goals from 24 hours to 4 hours is a measurable objective to enhance data resiliency.
Or getting the storage systems to deliver higher performance from 350 IOPS to 5000 IOPS for the database.
What I am saying here is these data points are measurable, and they can serve as checkpoints for business and operational improvements. From a management perspective, these can be used as KPI (key performance index) to define continuous improvement of Data Confidence.
Furthermore, it is easy when a OKR dashboard is used to map the improvement markers when organizations use storage to move from point A to point B, where B equates to a new success milestone. The alignment sets the paths to the business targets.
Storage does not mean only Capacity
The sad part is what the OKRs and the measured goals alignments are glaringly missing in the minds of many organizations purchasing a storage infrastructure and data management solution. The people tasked to source a storage technology solution are not placing a set of goals and objectives. Capacity appears to be the only thing on their mind.
I am about to meet a procurement officer of a customer soon. She asked me this question “Why is your storage so expensive?” over email. I want to change her mindset, just like the many officers and C-levels who hold the purse strings.
Let’s frame the use storage infrastructure in the real world. Nobody buys a storage system just to keep data in there much like a puddle keeps stagnant water. Sooner or later the value of the data in the storage evaporates or the value becomes dull if the data is not used well in any ways, shape or form.
Storage systems and the interconnected pathways from on premises, to the next premises, to the edge and to the clouds serve the greater good for Data. Data is used, shared, shaped, improved, enhanced, protected, moved, and more to deliver Value to the Business.
Storage capacity is just one of the few factors to consider when investing in a storage infrastructure solution. In fact, capacity is probably the least important piece when considering a storage solution to achieve the company’s OKRs. If we think about it deeper, setting the foundation for Data in the defensive manner will help elevate value of the data to be promoted with the offensive strategies to gain the competitive advantage.
Storage infrastructure and storage solutions along with data management platforms may appear to be a cost to the annual budgets. If you know set the OKRs, define A to get to B, alignment the goals, storage infrastructure and the data management platforms and practices are investments that are worth their weight in gold. That is my guarantee.
On the flip side, ignoring and avoiding OKRs, and set the strategies without prudence will yield its comeuppance. Technical debts will prevail.
On the road, seat belt saves lives. So does the motorcycle helmet. But these 2 technologies alone are probably not well received and well applied daily unless there is a strong ecosystem and culture about road safety. For decades, there have been constant and unrelenting efforts to enforce the habits of putting on the seat belt or the helmet. Statistics have shown they reduce road fatalities, but like I said, it is the safety culture that made all this happen.
On the digital front, the ransomware threats are unabated. In fact, despite organizations (and individuals), both large and small, being more aware of cyber-hygiene practices more than ever, the magnitude of ransomware attacks has multiplied. Threat actors still see weaknesses and gaps, and vulnerabilities in the digital realms, and thus, these are lucrative ventures that compliment the endeavours.
Time to look at Data Management
The Cost-Benefits-Risks Conundrum of Data Management
And I have said this before in the past. At a recent speaking engagement, I brought it up again. I said that ransomware is not a cybersecurity problem. Ransomware is a data management problem. I got blank stares from the crowd.
I get it. It is hard to convince people and companies to embrace a better data management culture. I think about the Cost-Benefits-Risk triangle while I was analyzing the lack of data management culture used in many organizations when combating ransomware.
I get it that Cybersecurity is big business. Even many of the storage guys I know wanted to jump into the cybersecurity bandwagon. Many of the data protection vendors are already mashing their solutions with a cybersecurity twist. That is where the opportunities are, and where the cool kids hang out. I get it.
Cybersecurity technologies are more tangible than data management. I get it when the C-suites like to show off shiny new cybersecurity “toys” because they are allowed to brag. Oh, my company has just implemented security brand XXX, and it’s so cool! They can’t be telling their golf buddies that they have a new data management culture, can they? What’s that?
Last week was consumed by many conversations on this topic. I was quite jaded, really. Unfortunately many still take a very simplistic view of all the storage technology, or should I say over-marketing of the storage technology. So much so that the end users make incredible assumptions of the benefits of a storage array or software defined storage platform or even cloud storage. And too often caveats of turning on a feature and tuning a configuration to the max are discarded or neglected. Regards for good storage and data management best practices? What’s that?
I share some of my thoughts handling conversations like these and try to set the right expectations rather than overhype a feature or a function in the data storage services.
Complex data networks and the storage services that serve it
I/O Characteristics
Applications and workloads (A&W) read and write from the data storage services platforms. These could be local DAS (direct access storage), network storage arrays in SAN and NAS, and now objects, or from cloud storage services. Regardless of structured or unstructured data, different A&Ws have different behavioural I/O patterns in accessing data from storage. Therefore storage has to be configured at best to match these patterns, so that it can perform optimally for these A&Ws. Without going into deep details, here are a few to think about:
Random and Sequential patterns
Block sizes of these A&Ws ranging from typically 4K to 1024K.
Causal effects of synchronous and asynchronous I/Os to and from the storage
“NVMe® and NVM®eF‰, as it evolves, can become the Great Peacemaker and bringing both divides and uniting them into a single storage fabric.“
I envisioned NVMe® and NVMe®oF™ setting the equilibrium at the storage architecture level, finishing the great storage fabric into one. This balance in the storage ecosystem at the storage interface specifications and language-protocol level has rapidly unifying storage today, and we are already seeing the end-to-end NVMe paths directly from the PCIe bus of one host to another, via networks over Ethernet (with RoCE, iWARP, and TCP flavours) and Fibre Channel™. Technically we can have an end point device, example a tablet, talking the same NVMe language to its embedded storage as well as a cloud NVMe storage in an exascale storage far, far away. In the past, there were just too many bridges, links, viaducts, aqueducts, bypasses, tunnels, flyovers to cross just to deliver a storage command, or a data in a formats, encased and encoded (and decoded) in so many different ways.
Colours in equilibrium, like the rainbow
Simple basics of NVMe®
SATA (Serial Attached ATA) and SAS (Serial Attached SCSI) are not optimized for solid state devices. besides legacy stuff like AHCI (Advanced Host Controller Interface) in SATA, and archaic SCSI-3 primitives in SAS, NVM® has so much to offer. It can achieve very high bandwidth and support 65,535 I/O queues, each with a queue depth of 65,535. The queue depth alone is a massive jump compared to SAS which has a queue depth limit of 256.
A big part of this is how NVMe® handles I/O processing. It has a submission queue (SQ) and a completion queue (CQ), and together they are know as a Queue Pair (QP). The NVMe® controller handles tens of thousands at I/Os (reads and writes) simultaneously, alerted to switch between each SQ and CQ very quickly using the MSI or MSI-X interrupt. Think of MSI and MSI-X as a service bell, a hardware register that informs the NVM® controller when there are requests in the SQ, and informs the hosts that there are completed requests in the CQ. There will be plenty of “dings” by the MSI-X service register but the NVMe® controller can perform it very well, with some smart interrupt coalescing.
NVMe I/O processing
NVMe® 1.1, as I recalled, used to be have 3 admin commands and 10 base commands, which made it very lightweight compared to SCSI-3. However, newer commands were added to NVMe® 2.0 specifications included command sets fo key-value operations and zoned named space.
Data movement is expensive. Not just costs, but also latency and resources as well. Thus there were many narratives to move compute closer to where the data is stored because moving compute is definitely more economical than moving data. I borrowed the analogy of the 2 animals from some old NetApp® slides which depicted storage as the elephant, and compute as birds. It was the perfect analogy, because the storage is heavy and compute is light.
“Close up of a white Great Egret perching on top of an African Elephant aa Amboseli national park, Kenya”
Before the animals representation came about I used to use the term “Data locality, Data Mobility“, because of past work on storage technology in the Oil & Gas subsurface data management pipeline.
Take stock of your data movement
I had recent conversations with an end user who has been paying a lot of dollars keeping their “backup” and “archive” in AWS Glacier. The S3 storage is cheap enough to hold several petabytes of data for years, because the IT folks said that the data in AWS Glacier are for “backup” and “archive”. I put both words in quotes because they were termed as “backup” and “archive” because of their enterprise practice. However, the face of their business is changing. They are in manufacturing, oil and gas downstream, and the definitions of “backup” and “archive” data has changed.
For one, there is a strong demand for reusing the past data for various reasons and these datasets have to be recalled from their cloud storage. Secondly, their data movement activities still mimicked what they did in the past during their enterprise storage days. It was a classic lift-and-shift when they moved to the cloud, and not taking stock of their data movements and the operations they ran on these datasets. Still ongoing, their monthly AWS cost a bomb.
We often hear the word “modernization” thrown around these days. The push is to get the end user to refresh their infrastructure, and the storage infrastructure market is rife with modernization word. Is your storage ripe for “modernization“?
Many possibilities to modernize storage
To modernize, it has to be relative to legacy storage hardware, and the operating environment that came with it. But if the so-called “legacy” still does the job, should you modernize?
Big Data is right
When the word “Big Data” came into prominence a while back, it stirred the IT industry into a frenzy. At one point, Apache Hadoop became the poster elephant (pun intended) for this exciting new segment. So many Vs came out, but I settled with 4 Vs as the framework of my IT conversations. The 4Vs we often hear are:
