sandrolowicz – Connect Worldwide

It’s here: Release 2 of HPE GreenLake for Block Storage powered by HPE Alletra Storage MP

sandrolowicz — Wed, 04 Oct 2023 21:51:10 +0000

Get higher performance, more scalability, and enhanced manageability with release 2 of HPE GreenLake for Block Storage built on HPE Alletra Storage MP.

–By Simon Watkins, Senior Manager, Worldwide Product Marketing for HPE Storage and Cloud Data Services

In April, HPE announced a major advancement in our HPE GreenLake edge-to-cloud platform: an expanded HPE GreenLake for Block Storage built on HPE Alletra Storage MP that provides mission-critical resiliency at midrange economics via the industry’s first disaggregated, scale-out block storage with a 100% data availability guarantee.[1]

Since launch, this new storage platform has enjoyed positive endorsements from industry analysts and press, as well as strong demand from partners and customers. That’s no surprise. HPE GreenLake for Block Storage removes many of the roadblocks and compromises associated with legacy storage infrastructure today. How? It’s a storage service that unlocks agility by delivering the cloud experience everywhere, maximizing infrastructure efficiency at scale, ensuring apps are always-on and always-fast, and eliminating price-performance compromises.

The introduction of an expanded HPE GreenLake for Block Storage powered by HPE Alletra Storage MP represented the first foundational step on a journey to consolidate and simplify our HPE block storage portfolio on a common software and hardware stack – all with a unified cloud management experience provided by the HPE GreenLake Cloud Platform.

Now for the next step on this “journey to one.” I’m excited to announce the introduction of a new and improved HPE GreenLake for Block Storage built on HPE Alletra Storage MP. With release 2, we’re delivering even more value for our customers with significant advancements in performance, scalability, connectivity, and manageability.

Let’s take a closer look at some of the key features of this new release and what it means for you.

What’s new in release 2?

You can:

Accelerate mission-critical apps with up to 2x performance boost

HPE GreenLake for Block Storage built on HPE Alletra Storage MP leverages a parallel, multi-node, all-active architecture with I/O processing active on all media, controllers, and host ports. This ensures consistent, predictable performance and ultra-low latency at scale. With release 2, we’re introducing new 32-core redundant controller nodes – adding to the existing choice of 8 core and 16 core redundant controller nodes and enabling a leap in performance of up to 2x.

Easily scale to meet future demands with 3x more capacity[2]

HPE GreenLake for Block Storage built on HPE Alletra Storage MP is designed to deliver disaggregated, scale-out storage with independent scaling of capacity and performance. Our initial offering at launch comprised a single 2U chassis with 24 NVMe SSD slots. Release 2 adds support for up to two JBOF expansion shelves, enabling you to start small and scale big – from 15.36TB to around 1.1 PB[3] – without disruption. And you can now maximize efficiency and cost savings with granular capacity upgrades. Support for 8 to 24 SSDs per enclosure in 2 drive increments gives you the flexibility to fine-tune, adapt, and elastically scale storage environments in line with specific workload capacity requirements and evolving business needs.

Deploy in ethernet environments with iSCSI host connectivity

Release 2 broadens your connectivity options with new support for iSCSI. You now have the flexibility of choosing the front-end storage protocol that best fits your workloads and existing network infrastructure with support for Fibre Channel, iSCSI, and NVMe-oF/FC.

Get more disaster recovery flexibility and efficiency via replication operability with HPE Primera and HPE Alletra 9000

HPE GreenLake for Block Storage replication software brings a rich set of features that can be used to design disaster-tolerant solutions that cost-effectively address disaster recovery challenges. It’s a uniquely easy, efficient, and flexible replication technology that allows you to protect and share data from any application. With release 2, you can now flexibly and cost-effectively configure replication partners to include HPE Primera, HPE Alletra 9000, or HPE GreenLake for Block Storage built on HPE Alletra Storage MP. You can replicate data over IP (using 10/25Gbe HBAs) or Fibre Channel (using 32/64 Gb HBAs) and meet any recovery-point objective (RPO) and recovery-time objective (RTO) with a choice of synchronous, synchronous long distanced, or asynchronous periodic data replication modes.

Secure a simple, affordable, risk-free transition

You have many reasons to move to the new HPE GreenLake for Block Storage platform. However, the process of transitioning from aging hardware is often fraught with challenges and concern. Now you can remove the risk and complexity from switching to a new storage platform via quick, non-disruptive data migration. With HPE Peer Motion,[4] you can move data from HPE 3PAR, HPE Primera, and HPE Alletra 9000 storage arrays to HPE GreenLake for Block Storage built on HPE Alletra Storage MP – online, non-disruptively, and without complex planning or dependency on extra tools.

Simplify management with enhanced and extended AI-driven performance reporting

HPE GreenLake for Block Storage built on HPE Alletra Storage MP is powered by industry leading AIOps for infrastructure, delivered via the Data Services Cloud Console (DSCC) on the HPE GreenLake platform. With release 2, we’re eliminating the guesswork for customers when it comes to storage optimization with enhanced and extended AI-based performance reporting and analytics for improved troubleshooting and insights.

Ensure investment protection with an expanded host OS integration ecosystem

Release 2 of HPE GreenLake for Block Storage built on HPE Alletra Storage MP adds Tier 2 Host OS Interoperability for Fibre Channel configurations as well as Tier 1 Host OS support for iSCSI environments. Consult SPOCK (Single Point of Connectivity Knowledge) for the latest interoperability status.

Ready to learn more?

Have a look at these resources:

And check out these videos:

What comes after development? Simplified deployment for HPE NonStop applications

sandrolowicz — Wed, 09 Aug 2023 22:27:26 +0000

What comes after development? Simplified deployment for HPE NonStop applications

Continuing the modern home analogy

What characterizes a home as modern was the feature of our previous post. Among the many items that are readily associated with modern homes are not just their open floor plans and their modern hardware but clear sight lines. The suggestion here is that homeowners today are looking for the ability to participate in all activities in the home, no matter where they stand. Homeowners want to feel connected and connection can only come through visibility.

The same could be said for modern IT development.

In our first post on the subject, HPE Mission Critical Systems Marketing Manager Vikas Kapoor drew the comparison between homeowners and what defines modernization of the development environment: “In many ways, IT development wants to leverage open technologies as much as possible even as they want connectivity to every corner of the enterprise along with the ease of movement that this entails, within the enterprise and beyond.”

Comparing today’s IT deployments with how modern homes function can only go so far and yet the imagery is hard to ignore.

Software developers today not only value development environments that are open with a vast array of functionality readily accessible and connected, but they also want to see how their application is deployed.

In many cases they want to see how the application adds value to an enterprise. A modern home might indeed be a family’s castle but for IT professionals, their IT systems are a fortress for their most value currency — data. Here is where every enterprise application ultimately ends up and where competition is defeated, authorized users are welcomed, and insights are gained and acted upon.

Modernizing application deployment on HPE NonStop

In deploying new applications it’s of particular importance that nothing should ever compromise the key NonStop attributes of availability, scalability and data integrity as oftentimes NonStop is the direct external interface. It is where touch points between the enterprise and users occur and where fresh data is created. NonStop is at the center of what matters most: data.

While data may be likened to gold, in reality it’s more like air. Without it, an enterprise cannot breathe. As fresh as data created on NonStop might be, without connections and ways to move data, its value diminishes rapidly with time.

Deployment? Modern aspects and requirements of applications to be deployed? Again, we can ask, what is modernization now and into the future?

HPE Master Technologist Justin Simonds says, “Modernization, when we used to speak about it, was platform based — modernize the interface (no SCOBOL), modernize the language to C or Java (no TAL/COBOL) and finally modernize the database. This ‘next generation’ of modernization is about the system fitting into the new cloud environment. This leads into DevOps, Git, Jenkins, etc. So, modernization at the company level rather than the platform level and not just a modern application, has more to do with fitting easily into the modern environment.”

Recognition of languages, databases, and the increased presence of cloud services is all part of the modernization equation that the NonStop community recognizes all too well. But modernization is more than just development. It’s just as essential to recognize the importance of deploying those applications in a standard environment with easy-to-use tools and methods.

IT professionals today know what they need to successfully develop an application. But when it comes time to deployment, the frameworks, visualization, and the way data is organized and then moved are often predetermined by enterprise IT.

There are well-defined expectations and for software developers to view any platform as being modern, conformity is unequivocal. The platform has to support those capabilities the enterprise has already deemed as being modern. For instance, when it comes to the world of hybrid IT, the need for consistency and conformity has never been greater. After all, the expectation is to use the tools you have and to capitalize on the investments already made.

“Developers no longer need to be concerned over the perceived intricacies of programming to a fault tolerant paradigm but rather all they need to be concerned about is their own programming efforts. Everything else will be looked after on their behalf,” says Kapoor. For instance, “Together, Pathsocket and NSJI libraries provided by the NonStop team help create TS/MP Serverclass applications without needing to know the specific APIs that make such applications special – the former for C/C++ and the latter for building Java based applications.”

NonStop with Java: Tools and Services to make deployment easy

When it comes to the makeup of a modern IT data center, NonStop professionals are fully aware of hybrid IT and the opportunities this provides. NonStop has become an active participant, either as a traditional server or virtually present in private clouds.

When it comes to the topic of Java and the many Java-centric services NonStop supports, this too is being seen as further evidence of the investment HPE is making in NonStop software. For the NonStop community this investment by HPE has been critical to NonStop media messages of modernization as Java is one of the preferred languages recognized as being modern. Java ensures a level of portability and in so supporting Java, NonStop can become host to many emerging solutions.

From humble beginnings from almost two decades past, today NonStop support for Java has reached a level where it is no longer just a case of porting an application to NonStop but having the tools to readily inherit NonStop fundamentals without having to acquire a deep understanding of the software modules unique to NonStop.

With what the Java ecosystem on NonStop (NSJ and its friends) now supports, there is no need for specialized NonStop skills to tackle the porting and then deployment of any Java-based mission critical applications. And to develop something from scratch, there are additional tools available as well. Regardless of whether porting or developing, deployment is made easier with the modern Java-based tools and supporting services available today on NonStop. When there is a commitment to DevOps, then these NonStop developers can be assured that the tools that they are familiar with are supported on NonStop.

What’s new with Java on NonStop

What is new with Java when it comes to deployment? How well does NonStop support the runtime environment Java applications require? With the availability of NSJ11 Update 2, the Java Virtual Machine and Java Runtime Environment offered on the HPE NonStop are now based on the OpenJDK 11 community version.

Any Java application that conforms to the Java standards can readily be deployed on the NSJ. This is an important milestone for NonStop as this brings an enterprise-class environment, fully leveraging the strength of a strong worldwide community, for server-side Java applications to NonStop. Just as important when it comes time to deploy a Java application that has been ported to suit the NonStop operating system and complies with the headless Java APIs defined in the standards. All this has been pursued to ensure the goal of making NonStop more open and modern can be readily recognized by all application developers.

HPE NonStop API Gateway and other connection enablers

Whether your application needs support for Enterprise Java Beans (EJB), a Tomcat compatible container, the latest HTTP server offering, or simply a gateway to other platforms, no anticipated application should stall at the time of deployment from anxieties over a lack of supporting infrastructure.

With NonStop Application Server for Java (NSASJ), NonStop Servlets for JavaServer Pages (NSJSP), a secure modern web server on NonStop (NSHTTP Server), and an API Gateway connecting NonStop to the rest of IT, all these bases are more than adequately covered. NonStop now has a track record of success with Java and Java development where today it represents the major development and deployment option for enterprise hybrid IT.

Deploying an application where there is no option to connect to other processes and databases would limit the usefulness of that application. The HPE NonStop API Gateway is provided to address these enterprise requirements. In addition to its functionality to connect NonStop to other platforms, this product participates in authentication and authorization, transforms data structures, does dynamic routing enabling canary releases of applications and other functionalities.

Looking further ahead, there are plans in place where NonStop API Gateway will offer service orchestration/aggregation, additional connectivity protocols, and more. For example, implementation of open interfaces where users can plug in their own custom subroutines and error codes is in an advanced stage of development.

However, it is the sustained interest in the provision of a REST/HTTP(S) interface carrying JSON formatted data payload and where it includes a built-in web server that is indicative of yet more functionality to come. Even as it already is capable of distributing/routing incoming requests to multiple services inside NonStop, the API Gateway also offers proxy functionality to send requests out of NonStop to remote servers.

Recently, HPE announced support for connectivity to the Apache Kafka messaging bus, too. The NonStop vendor community has already stepped into the Kafka arena with product offerings just as they offer products to support REST / HTTP(s). Major functionality of the API Gateway itself has been implemented using Apache Camel with the intent to provide an open-source framework that enables integration of NonStop with various other systems using several Enterprise Integration Patterns.

Develop and deploy with the mission critical leader

A modern home may indeed be defined as being open, with sight lines that lead to greater connectivity with the family even as they look for greater participation in all that happens around them. Much the same can be said about the modernization of IT.

Whether you look for your favorite development tools or simply want to ensure the infrastructure you need for deployment is present on NonStop rest assured the investments that HPE continues to make in NonStop means that today, they are all there for your benefit. Whether you develop in C/C++, Python or Java, what you need has been provided and in so doing, NonStop continues to be a leader in all things mission critical. The importance of NonStop? It is always there! The importance of data? It is the air that we breathe!

To learn more, visit www.hpe.com/info/nonstop

Meet the author:

Sridhar Neelakantan, Product Manager for HPE NonStop

Sridhar Neelakantan is a senior Product Manager for HPE NonStop and manages many products ranging from the NonStop middleware portfolio, languages and development environment with related cross-compliers, and the NonStop Development Environment for Eclipse (NSDEE). Sridhar is also the liaison from NonStop to the HPE Technology Partner Ready program. He has been with HPE and NonStop for more than 10 years, and works out of Bangalore, India.

HPE GreenLake for File Storage: Storage for Splunk’s cold tier that’s fast and simple

sandrolowicz — Wed, 09 Aug 2023 21:39:34 +0000

Discover the recognizable benefits of deploying Splunk in a cloud-enabled environment based on HPE GreenLake for File Storage.

–By Keith Vanderford, Storage Solutions Engineer, HPE

Splunk deployments typically have a high performance storage tier for hot and warm data, and a cheaper but slower storage tier for cold data. With this implementation, you have to balance search times against capacity constraints. Splunk’s searches are fastest when searching over recent data in your hot and warm buckets. For the best response times, it’s desirable to have all your data on that fastest tier of storage. But the capacity of that tier is usually limited due to the cost of fast flash-based storage. In order to increase capacity while trying to hold total cost down, organizations usually implement a second tier of storage for Splunk’s cold data, using less costly but slower storage technologies. Managing this second tier of storage introduces more complexity to your infrastructure, causing your staff to spend more time administering storage and less time extracting value from the data contained in that storage.

Questions?

I know you’ve got questions, like:

Shouldn’t my business be more about discovering valuable insights than managing storage?

Why does the second tier of storage for my cold data have to be slower?

Why does storage management have to be complicated?

HPE GreenLake for File Storage provides answers

Here’s good news: With HPE GreenLake for File Storage, your cold storage tier can be fast as well as simple and intuitive to manage.

HPE GreenLake for File Storage provides the perfect infrastructure for a cold storage tier that is both fast and easy to manage. It lets you take advantage of Splunk’s ability to provide the insights you need with fast searches over not just your hot and warm data, but your older (cold) data as well. HPE GreenLake for File Storage is an ultra-efficient all-NVMe storage solution with a cloud-like operational experience for data lakes. It delivers sustained, predictable throughput for enterprise performance at scale. The intuitive cloud interface also helps you reduce operational overhead.

You can reduce the performance penalty normally associated with searching older data by using the fast file-based storage provided by HPE GreenLake for File Storage for Splunk’s cold buckets. Simply mount NFS or SMB file shares provided by the ultra-efficient all-NVMe HPE GreenLake for File Storage solution to your Splunk indexers for cold buckets. These shares can also be used for Splunk’s frozen buckets if you have compliance or archive requirements. With this high performance storage solution, searches over Splunk’s cold buckets are extremely fast, accelerating search response times over traditional implementations that use slower storage for cold data.

The unmatched data reduction and low overhead data protection of HPE GreenLake for File Storage decrease the overall capacity required to store your data. For example, in our internal lab testing the observed data reduction rate has been about 3:1. This is significantly better than the reduction typically achieved for Splunk’s indexed data with most other storage platforms. Low overhead erasure coding is implemented using up to 146 data drives with 4 parity drives. This enables HPE GreenLake for File Storage to provide complete data protection with as little as 3% overhead. The combined benefits of this unique data reduction and low overhead data protection help make the most efficient use of your cold storage tier without slowing down searches over your older data.

How using HPE GreenLake for File Storage with your Splunk deployment makes setup and configuration simple

Creating the SMB or NFS file shares you need is quick and easy, and the self-service console gives you an intuitive cloud experience you can access from anywhere. This empowers you to free up your staff to work on adding value to your business rather than managing the day-to-day operations of your infrastructure.

HPE GreenLake for File Storage is available using a pay-as-you-go pricing model that gives you even more value for your infrastructure investment. You only pay for what you use, without having to pay for excess capacity. More resources are always at the ready to allow you to expand when you need to, but you never have to pay for them until you use them. Thus you can maximize the agility and value of your Splunk storage without the costs associated with overprovisioning.

Free your cold data from a slow storage tier and complicated infrastructure

With Splunk and HPE GreenLake for File Storage, you can have extremely fast searches over older data in your cold data tier, while simplifying the management of your storage. Get faster time to insights, and enable your data analysts and data scientists to unlock more value from your data.

To learn more, read the technical brief: Maximize your Splunk investment with HPE GreenLake for File Storage

Enhancing hybrid cloud data protection with HPE GreenLake for Backup and Recovery

sandrolowicz — Wed, 09 Aug 2023 21:04:09 +0000

Written by Ashwin Shetty of Hewlett Packard Enterprise

Discover the latest updates on availability and features to help organizations simplify hybrid cloud data protection with HPE GreenLake for Backup and Recovery.

As your organization leverages the cloud to grow your business, you start generating volumes of production data that must be efficiently protected as you scale. As you are aware, protecting and managing apps and data in a hybrid cloud can prove even more complex and costly than legacy on-prem solutions. You need to deal with:

Siloed data – This is the data that is difficult to access and increases cost and risk – while lowering efficiency. With the proliferation of data infrastructure, data is increasingly likely to be siloed by location, owner, data type, management platform, and more. Organizations must manage and protect multiple data silos in distributed locations and this level of complexity severely inhibits the innovation and business value that data offers. At the same time, it increases IT expense, risk of attack, and potential data loss. In response, IT leaders need platforms to unify data and eliminate siloed infrastructure.
Inconsistent protection policies – Data protection on-premises and across cloud tends to be inconsistently enforced and complex to manage. Often, no common protection policies create a corresponding increase in risk for your data. IT leaders need a solution that unifies management across hybrid cloud.
Growing cyberthreats – Ever-increasing ransomware attacks affect every industry. With data distributed across multiple locations on-premises and in the cloud, IT leaders need a solution to effectively protect their data against ransomware and malware threats.

How HPE GreenLake for Backup and Recovery solves these challenges

Agile, resilient data protection is key to always-on availability for data and applications in today’s dynamic hybrid cloud environments. While every organization has its own set of requirements, almost all focus on cost efficiency, simplicity, performance, scalability, and future-readiness when architecting a data protection strategy and evaluating new technologies.

HPE GreenLake for Backup and Recovery delivers the simplicity of the cloud experience while seamlessly protecting on-prem and cloud-native workloads. Delivered as a service, it leverages global protection policies to consistently protect workloads across hybrid cloud. In this way it eliminates data silos, multiple administrative touch points, and cumbersome point solutions.

Now I’m excited we excited to announce the availability of some key features:

Protection of Microsoft SQL Server database – HPE GreenLake for Backup and Recovery now protects Microsoft SQL Server databases running in VMware virtual machines and provides the same high efficiency protection as for VMware virtual machines. Application hosts are easily registered with the service and the hosted Microsoft SQL Server databases added to the HPE GreenLake for Backup and Recovery inventory where they can be selected for protection individually or as part of a protection group.
Protection of Amazon Relational Database Service – This service also enables snapshot orchestration of Amazon RDS instances within the customer’s AWS environment and provides rapid recovery from that snapshot. This new feature extends the cloud native backup of the service to protect your AWS resources.

Along with these new protection features, you can continue to protect on-premises VMware VMs and cloud-native workloads like Amazon EBS volumes and EC2 instances.

Enhancing HPE GreenLake for Backup and Recovery

HPE GreenLake for Backup and Recovery provides access to new features and versions as soon as they’re available. This service often upgrades features and adds functionality every month, without disrupting your work. This makes it so easy to protect your workloads when compared to traditional data protection approaches – where periodic upgrades are often so costly and disruptive that customers might wait months for the functionality in a new version (if they don’t choose to skip some upgrades altogether).

I’m excited to introduce the following features with HPE GreenLake for Backup and Recovery that were released recently:

Storage Volume Protection – Starting off with Storage Volume Protection, users can now protect any HPE array volume that is managed by Data Services Cloud Console (DSCC). Volume protection, using snapshots and replicas, can be configured using Data Ops Manager. This is an extension of our service, allowing you to back up your HPE arrays when you migrate to HPE GreenLake for Backup and Recovery.

Replication – This enhances the Protection Policies. You’ll now be able to replicate utilizing HPE StoreOnce Catalyst Copy from one on-premises destination to another on-premises location or to a Cloud Protection Store. Using replication, users will be able to efficiently keep multiple backup copies of their data in two different datacenters. Users who back up to on-premises protection stores as well as the Cloud Protection Store will be able to configure either of the on-premises backups to be the source of the transfer to the cloud. This extends on the 3-2-1 backup rule, by giving you the option to have your data in multiple locations. As part of these global protection policies, on-premises and cloud-native backups are all managed through one policy.

Dual authorization on HPE StoreOnce System – Users have the ability to configure dual authorization on their HPE StoreOnce systems to secure their data. This gives you the option to include more than one administrative user’s approval for any destructive actions against a given backup asset.

File and Folder recovery enhancements – Users can now recover individual files and folders from HPE StoreOnce Systems, and from the HPE Cloud Protection Store. Regardless of which backup destination is used, you can either browse directly to the files and folders to be restored, or search for them using the new search functionality.

Interested in learning more?

Watch this video on simplifying hybrid cloud protection with HPE GreenLake for Backup and Recovery:

And check out how in a few simple steps, you can find out the estimate cost savings of HPE GreenLake for Backup and Recovery when compared to other solutions. Use this link to access our free Cost Estimator Tool.

Ready to try the service?

You can experience the benefits of HPE GreenLake for Backup and Recovery— including all features and support — with a 90-day free evaluation.

New HPE Solutions with WEKA empower AI/ML and deep learning workloads

sandrolowicz — Tue, 11 Jul 2023 01:15:21 +0000

AI workloads demand extreme data pipeline performance. Learn how HPE Solutions with WEKA are architected to deliver that breakthrough performance.

–By Don Wilson, WEKA Partner Manager, HPE

With the shift toward data-first modernization, companies are anchoring their business strategies on data to gain a competitive advantage. From commercial enterprises to research organizations to public institutions, artificial intelligence (AI) is emerging as the key workload for maximizing data leverage.

AI-driven workloads are projected to grow at an 18% CAGR through 2025, with enterprises deploying them at an aggressive rate.* As generative AI, machine learning (ML), deep learning, and technical computing workloads move to the mainstream, your business and IT infrastructure must keep pace with new high-demand performance requirements

AI workloads bring new challenges

Typically, AI-driven workloads have different stages in a data pipeline that introduce very different I/O profiles. They also require the support of many small, random reads simultaneously against massively scalable data sets, high read and write throughput, high read and write IOPs and low latency. The different stages of AI workflows also have varying performance, scalability, availability, flexibility, metadata handling, and data management requirements. In response, enterprises end up using multiple storage siloes, copying data between them. This is inefficient and also can lead to delays in data processing, which can significantly reduce time to insight and value.

Until recently, much of the focus for AI has been on model development and GPU deployment, at times neglecting the importance of feeding data-hungry GPUs with sufficient data to train models. GPUs can become a wasted investment if left idle, waiting for data. The faster data reaches the GPUs, the faster you’ll get your business outcomes. For example, faster data performance can translate to faster insights. Faster insights allow your teams to react faster to trends, significantly improving customer experience and creating new revenue opportunities.

AI demands performance at scale that cannot be addressed with legacy storage

The performance requirement for next-generation workloads like AI data pipelines is complex – delivering low latency access while providing high throughput for large files and high IOPS for small files, often simultaneously at a massive scale. Often, legacy storage architectures are initially used to pilot AI workloads. However, as datasets grow these traditional storage architectures encounter scalability limitations and lack the performance characteristics required to meet production needs and expectations.

Legacy storage and file systems can become a bottleneck in data pipelines. Conversely, traditional distributed and parallel file systems are typically complex, requiring a team of expert admins to configure, adjust and maintain the solution. You need a holistic data platform approach to harness massive machine learning and deep learning datasets to deliver the performance and scale required to use them efficiently and effectively.

This is where HPE Solutions with WEKA enter the picture.

Don’t allow inadequate performance to become an impediment to fully utilizing AI

HPE Solutions with WEKA are built on the WEKA® Data Platform’s advanced architecture, which leverages NVMe-based flash storage and ultra-low latency, high-speed networking to deliver orders of magnitude improved performance for AI/ML and analytics workloads as well as technical computing applications that leverage high-performance computing (HPC) and GPU servers. HPE Solutions with WEKA take full advantage of the latest flash, networking, and computing technologies to provide customers with the highest-performing solution for AI, ML and HPC.

HPE Solutions with WEKA include the HPE Alletra 4110 data storage server, which complements the WEKA® Data Platform and is designed to run the most performance-demanding, data-centric workloads from data stores for machine learning, deep learning, and analytics. It achieves this by delivering an ideal balance of superior data capacity, throughput, IOPS, latency, and processing for performance-intensive workloads.

The symmetrical system design of the HPE Alletra 4110 data storage server evenly spreads resources across processors, configuring the right mix of resources to meet the workload requirements.

It boasts a datacenter resource efficient 1U standard rack depth design, supporting up to 20 new generation Enterprise and Data Center Standard Form Factor (EDSFF) NVMe SSDs to which up to 315GB/s of PCIe Gen5 bandwidth, providing the highest performing interconnect for a variety of peripherals. It supports multiple high-performance networking cards with as much as 800Gbps of raw networking bandwidth through the system, supporting both InfiniBand and Ethernet connectivity. HPE Alletra 4110 eliminates the costly compromises of storage appliances and general-purpose server infrastructure.

Performance without scale is an AI showstopper

Meeting capacity requirements is an important variable when evaluating potential performance. Machine Learning workloads often achieve better results from large amounts of training data. Fully harnessing the benefit of that data requires planning to enable performance at scale. Systems that perform adequately with smaller datasets (e.g., a few hundred terabytes) may not perform nearly as well when datasets reach petabyte capacity in the future.

The WEKA® Data Platform provides more than just speed – it also provides a highly scalable, POSIX-compliant file system. WEKA clusters built on the HPE Alletra 4110 data storage server can start as small as 46TB and scale to hundreds of petabytes in a single namespace.

WEKA® Data Platform is designed to transcend the limitations of legacy file systems that would leverage local storage, NFS, or block storage, making it ideal for data-intensive AI and HPC workloads. The performance of HPE Solutions with WEKA scales linearly as more servers are added to the storage cluster, so it can easily scale with the increasing business demands.

HPE Solutions with WEKA software deliver flexibility, accessibility, and ease of use

When consolidating storage systems intended to equip AI workloads and data pipelines, the data platform needs to support a wide spectrum of I/O profiles and workload requirements and different access methods since applications will require a variety of access protocols. The WEKA® Data Platform supports NFS, SMB, S3, and GPUDirect Storage and provides its own POSIX-compliant intelligent client while simultaneously accessing the same global namespace. A CSI plug-in for Kubernetes is also available for container-based, high-performance workload requirements.

In addition, the WEKA® Data Platform’s zero-tune architecture doesn’t require special configurations to support different data and performance profiles, making it dramatically easier to manage. HPE Solutions with WEKA offer not only ease of use but also simple scaling and seamless sharing of data in virtually any location, delivered through a single storage architecture that runs on-premises or in the hybrid cloud with the performance of all-flash arrays while also enabling the tiering of data to an external S3 Object Store to provide greater flexibility and improved economics.

Rest at ease knowing HPE Solutions with WEKA are fully tested and validated to work together

HPE and WEKA have developed and validated an optimized HPE Solution with WEKA based on the HPE Alletra 4110 data storage server and WEKA® Data Platform to deliver the highest possible performance at scale. In addition to the HPE Alletra 4110 data storage server, HPE also offers purpose-built systems for WEKA based on HPE ProLiant Gen10 Plus systems, including the HPE ProLiant DL325 Gen10 Plus V2 and the HPE ProLiant DL360 Gen10 Plus. The HPE ProLiant DL family of servers are flexible, reliable, and performance-optimized rack servers that also support flash NVMe, making them another potential platform that can be used to support WEKA. The HPE ProLiant Gen10 Plus systems offer flexible options, a versatile design, and increased levels of security.

A cloud experience delivered as a service

HPE Solutions with WEKA may be purchased and consumed as a service via the HPE GreenLake edge-to-cloud platform. HPE GreenLake for Compute Ops Management is provided standard with the HPE Alletra 4110. Depending on your business needs, this allows you to invest via traditional ownership or pay-per-use models. These new solutions reinforce HPE’s strategy of delivering a cloud operational experience that is simplified, automated, and spans edge to cloud.

“The new HPE Alletra 4110 data storage server, powered by WEKA’s industry-leading data platform software, is an excellent example of the innovation HPE and WEKA are delivering together for customers”, says Jeff Echols, Vice President of Channel and Strategic Partnerships at WEKA. “This new solution leverages best-in-class technologies like PCIe Gen 5 connected NVMe and low latency high-speed networking to deliver exceptional storage performance and capacity density in a more energy efficient package. It also reflects the value of a validated end-to-end solution that HPE and WEKA customers have come to depend on for the leading-edge capabilities, and we believe they will appreciate the solution’s performance, density scalability, and sustainability benefits.”

Meet Open Source enthusiast and Fpart project developer, Ganael Laplanche

sandrolowicz — Tue, 11 Jul 2023 00:21:41 +0000

As part of our blog series on open source experts, the HPE Developer team recently met up with Ganael Laplanche, the project developer for Fpart, a sysadmin-oriented tool that helps users sort files and pack them into bags or ‘partitions’. Here, we’ll introduce you to his work, how it came about, and learn more about what got Ganael involved with working with open source software.

Ganael, can you tell us a little about the tools Fpart and Fpsync?

The project started when I was working for a renowned center for biomedical research after a discussion with a friend of mine. We wanted to implement a fast bin-packing tool to produce filesystem tree partitions with the same size and number of files. The tool quickly evolved and got support for hooks that can be triggered when a partition is generated.

At that time, we needed to move petabyte-scale filesystems to freshly-acquired storage arrays. With its new hooking system, Fpart seemed to be a good basement to launch small migration jobs in parallel through our SSH cluster. Initial tests (see our article in French) were successful but we were still depending on our on-site scheduler to orchestrate submitted jobs and it was to be retired sooner or later. We needed a new scheduler.

That’s where Fpsync comes into play : the tool wraps Fpart and embeds its own scheduler to trigger small Rsync jobs to parallelize data migration by itself. It can leverage your SSH cluster to get the best from your data servers, acting as a powerful, standalone, data migration tool.

Of course, as an ardent open source supporter, those tools were released with an open source license (BSD 2-Clause “Simplified” License). They were quickly adopted by large companies (Intel, AWS, Microsoft, Alibaba, Oracle, …) as well as research centers to migrate petabyte-scale filesystems.

What attracted you to free software?

I first discovered free software by reading magazines that were surfing on Linux hype during mid-90’s (and trying their GNU/Linux distros offered on CDROM). But I really began to understand what free software meant later during my studies. I was immediately seduced by the thought that it exemplified humanity’s best attribute: the willingness to share knowledge in order to move forward together.

As a student, this was very important to me: it enabled me to learn more, as the code is freely available and the open source community very responsive. I quickly felt that I owed the community something in return; I didn’t want to use all that free software (as in free beer) without giving something back. So I started looking at how I could make my own contribution. This is where FreeBSD played a important role, acting as a catalyst…

Why did you come to FreeBSD as a development platform?

There are several reasons for that choice. As a curious student, I tried FreeBSD in the early 2000’s, testing version 4.5. What impressed me at that time was its documentation (“handbook”) and man (“manual”) pages. While GNU/Linux appeared complex to me, FreeBSD suddenly became more clear. With a very nice and welcoming community, it was the perfect platform for a newcomer into the UNIX world. I became hooked on FreeBSD and haven’t returned to any other system since.

I later came to understand another reason why FreeBSD appeared more clear. It is a homogeneous system, not a patchwork of very different projects. This makes a world of difference, as a specific version of FreeBSD represents a specific version of base components (called “world”) and kernel, offering up a complete system. World and kernel are all maintained by the same entity (FreeBSD developers) and, because of this, everything is consistent – from any options to the documentation and man pages. This delivers great value for users and guarantees a level of robustness and stability for the system.

FreeBSD is a good choice for developers because it is POSIX compliant. This is important if you want to produce portable code. Also, it is very easy to access source code for world, kernel and ports (third-party applications ported to FreeBSD). One can easily patch things and test the modifications, which is a bit harder on other systems where you would often have to install a dedicated source package to be able to patch it.

Finally, the system is a pleasure to administrate and update. I think I have not needed to reinstall my machine since the late 2000’s; I’ve only performed updates since. Third-party applications can now be easily installed and upgraded using binary packages, which was not the case when I first started using FreeBSD.

These are all the reasons why I use FreeBSD on my systems – not just for servers and development, but also as a daily desktop OS. Lots of people still think FreeBSD is not ready for everyday use on the desktop, but I am living proof that this is not true!

What other open source projects are you involved with?

I became a FreeBSD developer in 2011 and I now maintain more than 40 ports (A port is a set of patches and build options that makes a software work on FreeBSD. It also acts as the basis for binary packages). Maintaining ports is a fantastic hobby because on the one hand, you have the chance to work on your favorite OS, and on the other hand, you can contribute patches back upstream. This way, you are always connected with the different communities.

Aside from my FreeBSD activities, I have several personal projects. I mentioned Fpart and Fpsync, but I am also the author of ldapscripts, a set of tools used to simplify user and group management within an LDAP directory. They are quite old now, but they still do the job. I also worked on various smaller projects, such as sms1xxx kernel module (a port of Linux’ Siano DVB-T driver to FreeBSD, now deprecated in favor of webcamd), evtViewer (a viewer for Ms event log files) or Grpar (a Build engine group archive extract tool). I also wrote several courses (in French).

I also try to contribute to software I use when I find a bug (either by fixing it or at least by reporting it).

Is there anything else you’d like to share with our readers?

I owe a lot to free software. That’s mostly what allowed me to learn computing, making my career possible. That’s why I contribute back the most I can.

But that takes time (that is, personal time) and money (we need machines to test on, as well as power to run them). I am glad to see more and more companies supporting open source. Recently, HPE provided me with a replacement for my old server, I’ll never thank them enough for that kindness! This HPE ProLiant ML350 allows me to perform faster Fpart and Fpsync tests as well as compile code far more quicker than with my old machine. This is a sign that things are changing. I think everybody now understands why it is so important to support open source development. Providing hardware is a simple yet very efficient way of supporting open source developers, sharing code is another one. Let’s encourage companies to continue that way!

As for individuals, do not hesitate to report bugs or share code. You will participate in making great things and get fantastic feedback from the community!