Dell PowerFlex – For Business-Critical Workloads and AI
Wed, 21 Feb 2024 00:10:52 -0000
|Read Time: 0 minutes
AI—the buzzword that dances on the tongues of tech enthusiasts, executives, and coffee-break conversationalists alike. It's the shiny promise of automation, insights, and futuristic marvels. But let's step back from the AI dazzle for a moment. Beneath the glitz lies a fundamental truth: business-critical applications are the unsung heroes of organizational success. Enter Dell PowerFlex, the sturdy workhorse that ensures these applications run seamlessly.
The AI hype revisited
Imagine a room abuzz with anticipation. Faces lean forward, eager for the next AI revelation. If you've followed my previous blog, Can I Do That AI Thing on Dell PowerFlex, you know the answer. Yes, you can do that AI thing on PowerFlex. Being able to do AI shouldn’t be the end all be all for organizations. In fact, for most, it’s probably only a small portion of their IT operations. To that end, Dell PowerFlex isn't just built for AI. In fact, PowerFlex’s real strength isn’t AI at all.
Crushing the AI illusion
Let's peel back the layers. Dell PowerFlex isn't a mystical crystal ball, predicting stock market trends or composing poetry. Instead, it's the backbone supporting everyday business operations. Think databases, application servers, file servers—the workhorses that keep your organization humming. These workloads are the lifeblood of any enterprise, and their smooth functioning is non-negotiable. For many organizations, AI operations are a distant second. Why not optimize for the workhorses as well as prepare to support that new AI model?
The workload warriors
- Databases: Customer data, financial records, and inventory details all reside in databases. Dell PowerFlex ensures their availability, scalability, and performance.
- Application Servers: The engines behind web applications, APIs, and services. PowerFlex flexes its muscles here, providing the horsepower needed for user requests, transactions, and data processing.
- File Servers: Shared drives, document repositories, and collaboration spaces rely on file servers. PowerFlex ensures your files flow smoothly, whether you're sharing a presentation or collaborating on a project.
- And So Many Others: ERP systems, CRM platforms, virtual desktops—the list goes on. Each workload has its quirks, demands, and deadlines. Dell PowerFlex steps up, offering a unified platform that simplifies management and boosts performance.
Business-critical, Dell PowerFlex vital
These business-critical workloads are the heartbeat of organizations. They power customer interactions, financial transactions, and strategic decision-making. When these workloads hiccup, the entire operation feels it. That's where Dell PowerFlex shines. Its architecture leverages a robust and resilient software-defined storage (SDS) platform. Translation? It's agile, scalable, and resilient.
So, what's the secret sauce? PowerFlex leverages distributed storage resources, creating a pool of compute and storage nodes. These nodes collaborate harmoniously, distributing data and handling failures gracefully. Whether you're running a database query, serving up a web page, or analyzing mountains of data, PowerFlex ensures the show goes on.
The PowerFlex promise
Dell PowerFlex isn't just a hardware box—it's a promise. A promise to keep your workloads humming, your data secure, and your business thriving. So, the next time AI dazzles you with its potential, remember that PowerFlex is the sturdy engine of reliability in the background, ensuring the lights stay on, the servers stay responsive, and the wheels of progress keep turning.
In the grand scheme of IT, Dell PowerFlex takes center stage—an unassuming force that holds everything together. And as we navigate the AI landscape, let's tip our hats to the real heroes who keep the gears turning, one workload—AI included—at a time.
In the interest of full disclosure, this blog was created with the assistance of AI.
Author: Tony Foster
Twitter: @wonder_nerd
LinkedIn
Related Blog Posts
A Simple Poster at NVIDIA GTC – Running NVIDIA Riva on Red Hat OpenShift with Dell PowerFlex
Fri, 15 Mar 2024 21:45:09 -0000
|Read Time: 0 minutes
A few months back, Dell and NVIDIA released a validated design for running NVIDIA Riva on Red Hat OpenShift with Dell PowerFlex. A simple poster—nothing more, nothing less—yet it can unlock much more for your organization. This design shows the power of NVIDIA Riva and Dell PowerFlex to handle audio processing workloads.
What’s more, it will be showcased as part of the poster gallery at NVIDIA GTC this week in San Jose California. If you are at GTC, we strongly encourage you to join us during the Poster Reception from 4:00 to 6:00 PM. If you are unable to join us, you can view the poster online from the GTC website.
For those familiar with ASR, TTS, and NMT applications, you might be curious as to how we can synthesize these concepts into a simple poster. Read on to learn more.
NVIDIA Riva
For those not familiar with NVIDIA Riva, let’s start there.
NVIDIA Riva is an AI software development kit (SDK) for building conversational AI pipelines, enabling organizations to program AI into their speech and audio systems. It can be used as a smart assistant or even a note taker at your next meeting. Super cool, right?
Taking that up a notch, NVIDIA Riva lets you build fully customizable, real-time conversational AI pipelines, which is a fancy way of saying it allows you to process speech in a bunch of different ways including automatic speech recognition (ASR), text-to-speech (TTS), and neural machine translation (NMT) applications:
- Automatic speech recognition (ASR) – this is essentially dictation. Provide AI with a recording and get a transcript—a near perfect note keeper for your next meeting.
- Text-to-speech (TTS) – a computer reads what you type. In the past, this was often in a monotone voice. It’s been around for more than a couple of decades and has evolved rapidly with more fluid voices and emotion.
- Neural machine translation (NMT) – this is the translation of spoken language in near real-time to a different language. It is a fantastic tool for improving communication, which can go a long way in helping organizations extend business.
Each application is powerful in its own right, so think about what’s possible when we bring ASR, TTS, and NMT together, especially with an AI-backed system. Imagine having a technical support system that could triage support calls, sounded like you were talking to an actual support engineer, and could provide that support in multiple languages. In a word: ground-breaking.
NVIDIA Riva allows organizations to become more efficient in handling speech-based communications. When organizations become more efficient in one area, they can improve in other areas. This is why NVIDIA Riva is part of the NVIDIA AI Enterprise software platform, focusing on streamlining the development and deployment of production AI.
I make it all sound simple, however those creating large language models (LLMs) around multilingual speech and translation software know it’s not so. That’s why NVIDIA developed the Riva SDK.
The operating platform also plays a massive role in what can be done with workloads. Red Hat OpenShift enables AI speech recognition and inference with its robust container orchestration, microservices architecture, and strong security features. This allows workloads to scale to meet the needs of an organization. As the success of a project grows, so too must the project.
Why is Storage Important
You might be wondering how storage fits into all of this. That’s a great question. You’ll need high performance storage for NVIDIA Riva. After all, it’s designed to process and/or generate audio files and being able to do that in near real-time requires a highly performant, enterprise-grade storage system like Dell PowerFlex.
Additionally, AI workloads are becoming mainstream applications in the data center and should be able to run side by side with other mission critical workloads utilizing the same storage. I wrote about this in my Dell PowerFlex – For Business-Critical Workloads and AI blog.
At this point you might be curious how well NVIDIA Riva runs on Dell PowerFlex. That is what a majority of the poster is about.
ASR and TTS Performance
The Dell PowerFlex Solutions Engineering team did extensive testing using the LibriSpeech dev-clean dataset available from Open SLR. With this data set, they performed automatic speech recognition (ASR) testing using NVIDIA Riva. For each test, the stream was increased from 1 to 64, 128, 256, 384, and finally 512, as shown in the following graph.
The objective of these tests is to have the lowest latency with the highest throughput. Throughput is measured in RTFX, or the duration of audio transcribed divided by computation time. During these tests, the GPU utilization was approximately 48% without any PowerFlex storage bottlenecks. These results are comparable to NVIDIA’s own findings in in the NVIDIA Riva User Guide.
The Dell PowerFlex Solutions Engineering team went beyond just looking at how fast NVIDIA Riva could transcribe text, also exploring the speed at which it could convert text to speech (TTS). They validated this as well. Starting with a single stream, for each run the stream is changed to 4, 6, 8, and 10, as shown in the following graph.
Again, the goal is to have a low average latency with a high throughput. The throughput (RTFX) in this case is the duration of audio generated divided by computation time. As we can see, this results in a RTFX throughput of 391 with a latency of 91ms with ten streams. It is also worth noting that during testing, GPU utilization was approximately 82% with no storage bottlenecks.
This is a lot of data to pack into one poster. Luckily, the Dell PowerFlex Solutions Engineering team created a validated architecture that details how all of these results were achieved and how an organization could replicate them if needed.
Now, to put all this into perspective, with PowerFlex you can achieve great results on both spoken language coming into your organization and converting text to speech. Pair this capability with some other generative AI (genAI) tools, like NVIDIA NeMo, and you can create some ingenious systems for your organization.
For example, if an ASR model is paired with a large language model (LLM) for a help desk, users could ask it questions verbally, and—once it found the answers—it could use TTS to provide them with support. Think of what that could mean for organizations.
It's amazing how a simple poster can hold so much information and so many possibilities. If you’re interested in learning more about the research Dell PowerFlex has done with NVIDIA Riva, visit the Poster Reception at NVIDIA GTC on Monday, March 18th from 4:00 to 6:00 PM. If you are unable to join us at the poster reception, the poster will be on display throughout NVIDIA GTC. If you are unable to attend GTC, check out the white paper, and reach out to your Dell representative for more information.
Authors: Tony Foster | Twitter: @wonder_nerd | LinkedIn
Praphul Krottapalli
Kailas Goliwadekar
Using Dell PowerFlex and Google Distributed Cloud Virtual for Postgres Databases and How to Protect Them
Fri, 03 Nov 2023 23:27:04 -0000
|Read Time: 0 minutes
Did you know you can get the Google Cloud experience in your data center? Well now, you can! Using Google Distributed Cloud (GDC) Virtual and Dell PowerFlex enables the use of cloud and container workloads – such as Postgres databases – in your data center.
Looking beyond day one operations, the whole lifecycle must be considered, which includes assessing how to protect these cloud native workloads. That’s where Dell PowerProtect Data Manager comes in, allowing you to protect your workloads both in the data center and the cloud. PowerProtect Data Manager enhances data protection by discovering, managing, and sending data directly to the Dell PowerProtect DD series virtual appliance, resulting in unmatched efficiency, deduplication, performance, and scalability. Together with PowerProtect Data Manager, the PowerProtect DD is the ultimate cyber resilient data protection appliance.
In the following blog, we will unpack all this and more, giving you the opportunity to see how Dell PowerFlex and GDC Virtual can transform how you cloud.
What is Google Distributed Cloud Virtual?
We will start by looking at GDC Virtual and how it allows you to consume the cloud on your terms.
GDC Virtual provides you with a consistent platform for building and managing containerized applications across hybrid infrastructures and helps your developers become more productive across all environments. GDC Virtual provides all the mechanisms required to bring your code into production reliably, securely, and consistently while minimizing risk. GDC Virtual is built on open-source technologies pioneered by Google Cloud including Kubernetes and Istio, enabling consistency between cloud and on premises environments like PowerFlex. Anthos GKE (on GCP and on-prem), Anthos Service Mesh, and Anthos Config Management are the core building blocks of Anthos, which has integrations with platform-level services such as Stackdriver, Cloud Build, and Binary Authorization. GDC Virtual users purchase services and resources from the GCP Marketplace.
Figure 1. GDC Virtual components.
GDC Virtual puts all your IT resources into a consistent development, management, and control framework, automating low-value tasks across your PowerFlex and GCP infrastructure.
Within the context of GCP, the term ‘hybrid cloud’ describes a setup in which common or interconnected services are deployed across multiple computing environments, which include public cloud and on-premises. A hybrid cloud strategy allows you to extend the capacity and capabilities of your IT without the upfront capital expense investments of the public cloud while preserving your existing investments by adding one or more cloud deployments to your existing infrastructure. For more information, see Hybrid and Multi-Cloud Architecture Patterns.
PowerFlex delivers software defined storage to both virtual environments and bare metal hosts providing flexible consumption or resources. This enables both two-tier and three-tier architectures to match the needs of most any environment.
PowerFlex container storage
From the PowerFlex UI – shown in the following figure – you can easily monitor the performance and usage of your PowerFlex environment. Additionally, PowerFlex offers a container storage interface (CSI) and container storage modules (CSM) for integration with your container environment. The CSI/CSM allows containers to have persistent storage, which is important when working with workloads like databases that require it.
Figure 2. PowerFlex dashboard provides easy access to information.
To gain a deeper understanding of implementing GDC Virtual on Dell Powerflex, we invite you to explore our recently published reference architecture.
Dell engineers have recently prepared a PostgreSQL container environment deployed from the Google Cloud to a PowerFlex environment with GDC Virtual in anticipation of Kubecon. For those who have deployed Postgres from Google Cloud, you know it doesn’t take long to deploy. It took our team maybe 10 minutes, which makes it effortless to consume and integrate into workloads.
Once we had Postgres deployed, we proceeded to put it under load as we added records to it. To do this, we used pgbench, which is a built-in benchmarking tool in Postgres. This made it easy to fill a database with 10 million entries. We then used pgbench to simulate the load of 40 clients running 40 threads against the freshly loaded database.
Our goal wasn’t to capture performance numbers though. We just wanted to get a “warm” database created for some data protection work. That being said, what we saw on our modest cluster was impressive, with sub-millisecond latency and plenty of IO.
Data protection
With our containerized database warmed up, it was time to protect it. As you probably know, there are many ways to do this, some better than others. We’ll spend just a moment talking about two functional methods of data protection – crash consistent and application consistent backups. PowerProtect Data Manager supports both crash-consistent and application consistent database backups.
A “crash consistent” backup is exactly as the name implies. The backup application captures the volume in its running state and copies out the data regardless of what’s currently happening. It’s as if someone had just pulled the power cord on the workload. Needless to say, that’s not the most desirable backup state, but it’s still better than no backup at all.
That’s where an “application consistent” backup can be more desirable. An application consistent backup talks with the application and makes sure the data is all “flushed” and in a “clean” state prior to it being backed up. At least, that’s the simple version.
The longer version is that the backup application talks to the OS and application, asks them to flush their buffers – known as quiescing – and then triggers a snapshot of the volumes to be backed up. Once complete, the system then initiates a snapshot on the underlying storage – in this case PowerFlex – of the volumes used. Once the snapshots are completed, the application-level snapshots are released, the applications begin writing normally to it again, and the backup application begins to copy the storage snapshot to the protected location. All of this happens in a matter of seconds, many times even faster.
This is why application consistent backups are preferred. The backup can take about the same amount of time to run, but the data is in a known good state, which makes the chances of recovery much greater than crash consistent backups.
In our lab environment, we did this with PowerProtect Data Manager and PowerProtect DD Virtual Edition (DDVE). PowerProtect Data Manager provides a standardized way to quiesce a supported database, backup the data from that database, and then return the database to operation. This works great for protecting Kubernetes workloads running on PowerFlex. It’s able to create application consistent backups of the Postgres containers quickly and efficiently. This also works in concert with GDC Virtual, allowing for the containers to be registered and restored into the cloud environment.
Figure 3. An application consistent backup and its timing in the PowerProtect Data Manager UI
It’s great having application consistent backups of your cloud workloads, “checking” many of those boxes that people require from their backup environments. That said, just as important and not to be forgotten is the recovery of the backups.
Data recovery
As has been said many times, “never trust a backup that hasn’t been tested.” It’s important to test any and all backups to make sure they can be recovered. Testing the recovery of a Postgres database running in GDC Virtual on PowerFlex is as straightforward as can be.
The high-level steps are:
- From the PowerProtect Data Manager UI, select Restore > Assets, and select the Kubernetes tab. Select the checkbox next to the protected namespace and click Restore.
- On the Select Copy page, select the copy you wish to restore from.
- On the Restore Type page, select where it should be restored to.
- Determine how the Persistent Volume Claims (PVCs) and namespace should be restored.
- When finished, test the restore.
You might have noticed in step 4, I mentioned PVCs, which are the container’s connections to the data and, as the name implies, allow that data to persist across the nodes. This is made possible by the CSI/CSM mentioned earlier. Because of the integration across the environment, restoring PVCs is a simple task.
The following shows some of the recovery options in PowerProtect Data Manager for PVCs.
Figure 4. PowerProtect Data Manager UI – Namespace restore options
The recovery, like most things in data protection, is relatively anticlimactic. Everything is functional, and queries work as expected against the Postgres database instance.
Dell and Google Cloud collaborated extensively to create solutions that leverage both PowerFlex and GDC Virtual. The power of this collaboration really shows through when recovery operations just work. That consistency and ease enables customers to take advantage of a robust environment backed by leaders in the space and helps to remove one nightmare that keeps developers and IT admins awake at night, allowing them to rest easy and be prepared to change the world.
If any of this sounds interesting to you and you’ll be at Kubecon in Chicago, Illinois on November 6-9, stop by the Google Cloud booth. We’ll be happy to show you demos of this exciting collaboration in action. Otherwise, feel free contact your Dell representative for more details.
Resources
Authors:
Authors: | Tony Foster, | Vinod Kumar Kumaresan, | Harsha Yadappanavar, |
LinkedIn: | |||
X (formerly Twitter): |
| @harshauy | |
Personal Blog: |
|
|