Category Archives: Cloud Storage
Broadcom server-storage connectivity sales down but recovery coming Blocks and Files – Blocks and Files
Although Broadcom saw an overall rise in revenues and profit in its latest quarter, sales in the server-to-storage connectivity area were down. It expects a recovery and has cash for an acquisition.
Revenues in Broadcoms third fiscal 2021 quarter, ended August 1, were $6.78 billion, up 16 per cent on the year. There was a $1.88 billion profit, more than doubling last years $688 million.
Were interested because Broadcom makes server-storage connectivity products such as Brocade host bus adapters (HBAs), SAS and NVMe connectivity products.
President and CEO Hock Tans announcement statement said: Broadcom delivered record revenues in the third quarter reflecting our product and technology leadership across multiple secular growth marketsin cloud, 5G infrastructure, broadband,and wireless. We are projecting the momentum to continue in the fourth quarter.
There are two segments to its business: Semiconductor Solutions, which brought in $5.02 billion, up 19 per cent on the year; and Infrastructure Software, which reported $1.76 billion, an increase of ten per cent.
Tan said in the earnings call: Demand continued to be strong from hyper-cloud and service provider customers. Wireless continued to have a strong year-on-year compare. And while enterprise has been on a trajectory of recovery, we believe Q3 is still early in that cycle, and that enterprise was down year on year.
Inside Semiconductor Solutions, the server storage connectivity area had revenues of $673 million, which was nine per cent down on the year-ago quarter. Tan noted: Within this, Brocade grew 27 per cent year on year, driven by the launch of new Gen 7 Fibre Channel SAN products.
Overall, Tan said: Our [Infrastructure Solutions] products here supply mission-critical applications largely to enterprise, which, as I said earlier, was in a state of recovery. That being said, we have seen a very strong booking trajectory from traditional enterprise customers within this segment. We expect such enterprise recovery in server storage.
This will come from aggressive migration in cloud to 18TB disk drives and a transition to next-generation SAS and NVMe products. Tan expects Q4 server storage connectivity revenue to be up low double-digit percentage year on year. Think two to five per cent.
The enterprise segment will grow more, with Tan saying: Because of strong bookings that we have been seeing now for the last three months, at least from enterprise, which is going through largely on the large OEMs, who particularly integrate the products and sell it to end users, we are going to likely expect enterprise to grow double digits year on year in Q4.
That enterprise business growth should continue throughout 2022, Tan believes: In fact, I would say that the engine for growth for our semiconductor business in 2022 will likely be enterprise spending, whether its coming from networking, one sector for us, and/or from server storage, which is largely enterprise, we see both this showing strong growth as we go into 2022.
Broadcom is accumulating cash and could make an acquisition or indulge in more share buybacks. Tan said: By the end of October, our fiscal year, well probably see the cash net of dividends and our cash pool to be up to close to $13 billion, which is something like $6 billion, $7 billion, $8 billion above what we would, otherwise, like to carry on our books.
Let us pronounce that HBAs are NICs (Network Interface Cards) and that an era of SmartNICs is starting. It might be that Broadcom could have an acquisitive interest in the SmartNIC area.
Broadcom is already participating in the DPU (Data Processing Unit) market, developing and shipping specialised silicon engines to drive specialised workloads for hyperscalers. Answering an analyst question, Tan said: We have the scale. We have a lot of the IP calls and the capability to do all those chips for those multiple hyperscalers who can afford and are willing to push the envelope on specialised I used to call it offload computing engines, be they video transcoding, machine learning, even what people call DPUs, smart NICs, otherwise called, and various other specialised engines and security hardware that we put in place in multiple cloud guys.
Better add Broadcom to the list of DPU vendors such as Fungible, Intel and Pensando, and watch out for any SmartNIC acquisition interest.
The rest is here:
Broadcom server-storage connectivity sales down but recovery coming Blocks and Files - Blocks and Files
"Rockset is on a mission to deliver fast and flexible real-time analytics" – JAXenter
JAXenter: Thank you for taking the time to speak with us! Can you tell us more about Rockset and how it works? How does it help us achieve real-time analytics?
Venkat Venkataramani: Rockset is a real-time analytics database that serves low latency applications. Think real-time logistics tracking, personalized experiences, anomaly detection and more.Wh
Rockset employs the same indexing approach used by the systems behind the Facebook News Feed and Google Search, which were built to make data retrieval for millions of users and on TBs of data, instantaneous. It goes a step further by building a Converged Index a search index, a columnar store and a row index on all data. This means sub-second search, aggregations and joins without any performance engineering.
You can point Rockset at any data structured, semi-structured and time series data and it will index the data in real-time and enable fast SQL analytics. This frees teams from time-consuming and inflexible data preparation. Teams can now onboard new datasets and run new experiments without being constrained by data operations. And, Rockset is fully-managed and cloud-native, making a massively distributed real-time data platform accessible to all.
SEE ALSO: Shifting toward more meaningful insights means shifting toward proactive analytics
JAXenter: What data sources does it currently support?
Venkat Venkataramani: Rockset has built-in data connectors to data streams, OLTP databases and data lakes. These connectors are all fully-managed and stay in sync with the latest data. That means you can run millisecond-latency SQL queries within 2 seconds of data being generated. Rockset has built-in connectors to Amazon DynamoDB, MongoDB, Apache Kafka, Amazon Kinesis, PostgreSQL, MySQL, Amazon S3 and Google Cloud Storage. Rockset also has a Write API to ingest and index data from other sources.
JAXenter: Whats new at Rockset and how will it continue to improve analytics for streaming data?
Venkat Venkataramani: We recently announced a series of product releases to make real-time analytics on streaming data affordable and accessible. With this launch, teams can use SQL to transform and pre-aggregate data in real-time from Apache Kafka, Amazon Kinesis and more.
This makes real-time analytics up to 100X more cost-effective on streaming data. And, we free engineering teams from needing to construct and manage complex data pipelines to onboard new streaming data and experiment on queries. Heres what weve released:
You can delve further into this release by watching a live Q&A with Tudor Bosman, Rocksets Chief Architect. He delves into how we support complex aggregations on rolled up data and ensure accuracy even in the face of dupes and latecomers.
JAXenter: What are some common use cases for real-time data analytics? When is it useful to implement?
Venkat Venkataramani: You experience real-time analytics every day whether you realize it or not. The content displayed in Instagram newsfeeds, the personalized recommendations on Amazon and the promotional offers from Uber Eats are all examples of real-time analytics. Real-time analytics encourages users to take desired actions from reading more content, to adding items to our cart, to using takeout and delivery services for more of our meals.
We think real-time analytics isnt just useful to the big tech giants. Its useful across all technology companies to drive faster time to insight and build engaging experiences. Were seeing SaaS companies in the logistics space provide real-time visibility into the end-to-end supply chain, route shipments and predict ETAs. This ensures that materials arrive on time and within schedule, even in the face of an increasingly complex chain. Or, there are marketing analytics software companies that need to unify data across a number of interaction points to create a single view of the customer. This view is then used for segmentation, personalization and automation of different actions to create more compelling customer experiences.
Theres a big misperception in the space that a) real-time analytics is too expensive b) real-time analytics is only accessible to large tech companies. Thats just not true anymore. The cloud offerings, availability of real-time data and the changing resource economics are making this within reach of any digital disrupter.
JAXenter: How is Rockset built under the hood?
Venkat Venkataramani: The Converged Index, mentioned previously, is the key component in enabling real-time analytics. Rockset stores all its data in the search, column-based and row-based index structures that are part of the Converged Index, and so we have to ensure that the underlying storage can handle both reads and writes efficiently. To meet this requirement, Rockset uses RocksDB as its embedded storage engine, with some modifications for use in the cloud. RocksDB enables Rockset to handle high write rates, leverage SSDs for optimal price-performance and support updates to any field.
Another core part of Rocksets design is its use of a disaggregated architecture to maximize resource efficiency. We use an Aggregator-Leaf-Tailer (ALT) architecture, common at companies like Facebook and LinkedIn, where resources for ingest compute, query compute and storage can be scaled independently of each other based on the workload in the system. This allows Rockset users to exploit cloud efficiencies to the full.
SEE ALSO: Codespaces helps developers to focus on what matters mostbuilding awesome things
JAXenter: Personally, what are some of your favorite open source tools that you cant do without?
Venkat Venkataramani: RocksDB! The team at Rockset built and open-sourced RocksDB at Facebook, a high performance embedded storage engine used by other modern data stores like CockroachDB, Kafka and Flink. RocksDB was a project at Facebook that abstracted access to local stable storage so that developers could focus their energies on building out other aspects of the system. RocksDB has been used at Facebook as the embedded storage for spam detection, graph search and message queuing. At Rockset, weve continued to contribute to the project as well as release RocksDB-cloud to the community.
We are also fans of the dbt community, an open-source tool that lets data teams collaborate on transforming data in their database to ship higher quality data sets, faster. We share a similar outlook on the data space we think data pipelines are challenging to build and maintain, respect SQL as the lingua franca of analytics and want to make it easy for data to be shared across an organization.
JAXenter: Can you share anything about Rocksets future? Whats on the roadmap next, what features and/or improvements are being worked on?
Venkat Venkataramani: Rockset is on a mission to deliver fast and flexible real-time analytics, without the cost and complexity. Our product roadmap is geared towards enabling all digital disrupters to realize real-time analytics.
This requires taking steps to make real-time analytics more affordable and accessible than ever before. A first step towards affordability was the release of SQL-based rollups and transformations, which cut the cost of real-time analytics up to 100X for streaming data. As part of our expansion initiative, were also expanding Rockset to users across the globe. Follow us as we continue to put real-time analytics within reach of all engineers.
Read the original:
"Rockset is on a mission to deliver fast and flexible real-time analytics" - JAXenter
Lessons Learned: Training and Deploying State of the Art Transformer Models at Digits – insideBIGDATA
In this blog post, we want to provide a peek behind the curtains on how we extract information with Natural Language Processing (NLP). Youll learn how to appy state-of-the-art Transformer models for this problem and how to go from an ML model idea to integration in the Digits app.
Our Plan
Information can be extracted from unstructured text through a process called Named Entity Recognition (NER). This NLP concept has been around for many years, and its goal is to classify tokens into predefined categories, such as dates, persons, locations, and entities.
For example, the transaction below could be transformed into the following structured format:
We had seen outstanding results from NER implementations applied to other industries and we were eager to implement our own banking-related NER model. Rather than adopting a pre-trained NER model, we envisioned a model built with a minimal number of dependencies. That avenue would allow us to continuously update the model while remaining in control of all moving parts. With this in mind, we discarded available tools like the SpaCy NER implementation or HuggingFace models for NER. We ended up building our internal NER model based only on TensorFlow 2.x and the ecosystem library TensorFlow Text.
The Data
Every Machine Learning project starts with the data, and so did this one. We decided which relevant information we wanted to extract (e.g., location, website URLs, party names, etc.) and, in the absence of an existing public data set, we decided to annotate the data ourselves.
There are a number of commercial and open-source tools available for data annotation, including:
The optimal tool varies with each project, and is a question of cost, speed, and useful UI. For this project, our key driver for our tool selection was the quality of the UI and the speed of the sample processing, and we chose doccano.
At least one human reviewer then evaluated each selected transaction, and that person would mark the relevant sub-strings as shown above. The end-product of this processing step was a data set of annotated transactions together with the start- and end-character of each entity within the string.
Selecting an Architecture
While NER models can also be based on statistical methods, we established our NER models on an ML architecture called Transformers. This decision was based on two major factors:
The initial attention-based model architecture was the Bidirectional Encoder Representation from Transformers (BERT, for short), published in 2019. In the original paper by Google AI, the author already highlighted potential applications to NER, which gave us confidence that our transformer approach might work.
Furthermore, we had previously implemented various other deep-learning applications based on BERT architectures and we were able to reuse our existing shared libraries. This allowed us to develop a prototype in a short amount of time.
BERT models can be used as pre-trained models, which are initially trained on multi-lingual corpi on two general tasks: predicting mask tokens and predicting if the next sentence has a connection to the previous one. Such general training creates a general language understanding within the model. The pre-trained models are provided by various companies, for example, by Google via TensorFlow Hub. The pre-trained model can then be fine-tuned during a task-specific training phase. This requires less computational resources than training a model from scratch.
The BERT architecture can compute up to 512 tokens simultaneously. BERT requires WordPiece tokenization which splits words and sentences into frequent word chunks. The following example sentence would be tokenized as follows:
Digits builds a real-time engine
[bdig, b##its, bbuilds, ba, breal, b-, btime, bengine]
There are a variety of pre-trained BERT models available online, but each has a different focus. Some models are language-specific (e.g., CamemBERT for French or Beto for Spanish), and other models have been reduced in their size through model distillation or pruning (e.g., ALBERT or DistilBERT).
Time to Prototype
Our prototype model was designed to classify the sequence of tokens which represent the transaction in question. We converted the annotated data into a sequence of labels that matched the number of tokens generated from the transactions for the training. Then, we trained the model to classify each token label:
In the figure above, you notice the O tokens. Such tokens represent irrelevant tokens, and we trained the classifier to detect those as well.
The prototype model helped us demonstrate a business fit of the ML solution before engaging in the full model integration. At Digits, we develop our prototypes in GPU-backed Jupyter notebooks. Such a process helps us to iterate quickly. Then, once we confirm a business use-case for the model, we focus on the model integration and the automation of the model version updates via our MLOps pipelines.
Moving to Production
In general, we use TensorFlow Extended (TFX) to update our model versions. In this step, we convert the notebook code into TensorFlow Ops, and here we converted our prototype data preprocessing steps into TensorFlow Transform Ops. This extra step allows us later to train our model versions effectively, avoid training-serving skew, and furthermore allows us to bake our internal business logic into our ML models. This last benefit helps us to reduce the dependencies between our ML models and our data pipeline or back-end integrations.
We are running our TFX pipelines on Google Clouds Vertex AI pipelines. This managed service frees us from maintaining a Kubernetes cluster for Kubeflow Pipelines (which we have done prior to using Vertex AI).
Our production models are stored in Google Cloud Storage buckets, and TFServing allows us to load model versions directly from cloud storage. Because of the dynamic loading of the model versions, we dont need to build custom containers for our model serving setup; we can use the pre-built images from the TensorFlow team.
Here is a minimal setup for Kubernetes deployment:
apiVersion: apps/v1 kind: Deployment metadata: name: tensorflow-serving-deployment spec: template: spec: containers: name: tensorflow-serving-container image: tensorflow/serving:2.5.1 command: /usr/local/bin/tensorflow_model_server args: port=8500 model_config_file=/serving/models/config/models.conf file_system_poll_wait_seconds=120
Note the additional argument file_system_poll_wait_seconds in the list above. By default, TFServing will check the file system for new model versions every 2s. This can generate large Cloud Storage costs since every check triggers a list operation, and storage costs are billed based on the used network volume. For most applications, it is fine to reduce the file system check to every 2 minutes (set the value to 120 seconds) or disable it entirely (set the value to 0).
For maintainability, we keep all model-specific configurations in a specific ConfigMap. The generated file is then consumed by TFServing on boot-up.
apiVersion: v1 kind: ConfigMap metadata: namespace: ml-deployments name:
After the initial deployment, we started iterating to optimize the model architecture for high throughput and low latency results. This meant optimizing our deployment setup for BERT-like architectures and optimizing the trained BERT models. For example, we optimized the integration between our data processing Dataflow jobs and our ML deployments, and shared our approach in our recent talk at the Apache Beam Summit 2021.
Results
The deployed NER model allows us to extract a multitude of information from unstructured text and make it available through Digits Search.
Here are some examples of our NER model extractions:
The Final Product
At Digits, an ML model is never itself the final product. We strive to delight our customers with well-designed experiences that are tightly integrated with ML models, and only then do we witness the final product. Many additional factors come into play:
Latency vs. Accuracy
A more recent pre-trained model (e.g., BART or T5) could have provided higher model accuracy, but it would have also increased the model latency substantially. Since we are processing millions of transactions daily, it became clear that model latency is critical for us. Therefore, we spent a significant amount of time on the optimization of our trained models.
Design for false-positive scenarios
There will always be false positives, regardless of how stunning the model accuracy was pre-model deployment. Product design efforts that focus on communicating ML-predicted results to end-users are critical. At Digits, this is especially important because we cannot risk customers confidence in how Digits is handling their financial data.
Automation of model deployments
The investment in our automated model deployment setup helped us provide model rollback support. All changes to deployed models are version controlled, and deployments are automatically executed from our CI/CD system. This provides a consistent and transparent deployment workflow for our engineering team.
Devise a versioning strategy for release and rollback
To assist smooth model rollout and a holistic quantitative analysis prior to rollout, we deploy two versions of the same ML model and use TFServings version labels (e.g., release and pre-release tags) to differentiate between them. Additionally, we use an active version table that allows for version rollbacks, made as simple as updating a database record.
Assist customers, dont alienate them
Last but not least, the goal for our ML models should always be to assist our customers in their tasks instead of alienating them. That means our goal is not to replace humans or their functions, but to help our customers with cumbersome tasks. Instead of asking people to extract information manually from every transaction, well assist our customers by pre-filling extracted vendors, but they will always stay in control. If we make a mistake, Digits makes it easy to overwrite our suggestions. In fact, we will learn from our mistakes and update our ML models accordingly.
Further Reading
Check out these great resources for even more on NER and Transformer models:
About the Author
Hannes Hapke is a Machine Learning Engineer at Digits. As a Google Developer Expert, Hannes has co-authored two machine learning publications: NLP in Action by Manning Publishing, and Building Machine Learning Pipelines by OReilly Media. At Digits, he focuses on ML engineering and applies his experience in NLP to advance the understanding of financial transactions.
Sign up for the free insideBIGDATAnewsletter.
Join us on Twitter:@InsideBigData1 https://twitter.com/InsideBigData1
See original here:
Lessons Learned: Training and Deploying State of the Art Transformer Models at Digits - insideBIGDATA
Save big on this decentralized cloud storage system | TheHill – The Hill
The Hill may be compensated and/or receive an affiliate commission if you buy through our links.
Safely storing data is a must for everyone, and for the security conscious, zero-knowledge cloud file storage is the ultimate way to go. But, what exactly is zero-knowledge file storage? We thought you would never ask. Zero-knowledge encryption means you are the only one who has the keys to access your data. It keeps your data private and for your eyes only. This means you and only you can retrieve your files, and you control file sharing right down to the intricate details.
Internxt Drive is one such cloud-based end-to-end encryption zero-knowledge file storage service. Typically, a year subscription for 2TB of decentralized cloud storage would cost $126, but it's on sale now for just $9.99.
Internxt Drive protects your privacy and offers the best in security. It works well for personal use and in a team environment, so you can securely store and share data. How does it work? Each file you upload to Internxt is client-side encrypted and divided into fragments. You are the only one who holds the key to said data, and that means you are the only one who can access your files.
There are a few zero-knowledge options out there, and Internxt Drive is one that comes with stellar reviews. TechRadar wrote, ""Unlike popular cloud storage services like Google Drive, Dropbox, and Microsoft OneDrive, Internxt is a zero-knowledge file storage service that supports end-to-end encryption."
Compatible with Google Drive, Microsoft OneDrive, Dropbox, and Apple iCloud means you can work the way you always have, but with uncompromising security. Collaborate privately with user-to-user solution, customizations, and featuresit really is the perfect solution for the security conscious. The intuitive interface is user friendly and offers easy to follow UI so you can easily store data. Internxt is available on all of your devices, and that means you can access your files, documents, videos, and more from the desktop app, web browser app, and the iOS/Android app.
Get the Internxt - 2TB Decentralized Cloud Storage: 1-Year Subscription for $9.99 (reg. $126), a discount of 92%.
Prices subject to change.
Here is the original post:
Save big on this decentralized cloud storage system | TheHill - The Hill
AWS, NetApp unveil the first fully-managed ONTAP file system in the cloud – ZDNet
Amazon
Amazon Web Services and NetApp on Thursday rolled out a new storage service years in the making, Amazon FSx for NetApp ONTAP. It's the first fully-managed ONTAP file system in the cloud, giving customers a simple way to run their applications on AWS without changing their code or how they manage data.
The new service comes nearly three years after AWS introduced the first iterations of FSx for Windows File Server and Lustre. FSx lets customers choose the file system that powers their file storage, giving them full access to the file system's feature sets and performance -- it runs fully-managed in the cloud. Customers get the benefits of AWS performance, along with seamless integration with other AWS services.
ONTAP is NetApp's file system technology that has traditionally powered network-attached storage (NAS) on premise. NAS is a large market with a number of major players like NetApp, Dell Technologies, HPE, Pure Storage and Nutanix, to name a few. Most enterprise applications rely on NAS, so multiple servers running the application can access a shared dataset.
For instance, critical programs like Epic's electronic health record (EHR) software use NAS to give doctors within a hospital access to a system of record. Typically, that application physically sits in a hospital. Now, it can live on AWS.
"This opens up an enormous opportunity for both of us and our customers to really enable more and more digital transformation and adoption of AWS," Anthony Lye, EVP and GM of public cloud at NetApp, said to ZDNet.
Lye said it's "equally great" for their competitors' customers, whom they plan to scoop up with this new offering. "We think we can take a lot of [market] share," he said.
"If you're running EMC and you rely on EMC storage on premise, you haven't had a path to the cloud," Lye continued. "You'd have to re-factor, you'd have to re-engineer. ONTAP is the gold standard, and EMC customers know and appreciate ONTAP. They now have confidence that ONTAP is a first-party Amazon service, and we're going to tell all those customers."
The service will serve customers that simply want to stop managing storage on premises.
It will also help customers that have hybrid goals, such as bursting to the cloud, maintaining business continuity or dev testing.
There are no upfront commitments or costs to use Amazon FSx for NetApp ONTAP, and customers only pay for the resources used.
Amazon FSx for NetApp ONTAP is available globally right out of the gate. It builds on nine years of partnership between Amazon and NetApp.
View original post here:
AWS, NetApp unveil the first fully-managed ONTAP file system in the cloud - ZDNet
Backblaze teams up with Vultr for new cloud storage competitor to Amazon, Google, and Microsoft – 9to5Mac
Many people think of Backblaze as just a cloud backup provider for Mac users looking to have an offsite copy of their photos, movies, documents, etc. The company also offers some robust business solutions as well. Today, Backblaze is announcing a new Backblaze B2 partnership with Vultr that gives organizations a storage alternative to the big three technology companies.
Scalable cloud storage and compute are necessities for most modern applications, said Nilay Patel, VP of Sales and Partnerships at Backblaze. With Backblaze and Vultr together, its no longer a necessity for organizations to tolerate vendor lock-in, complexities, and costs that have traditionally come with legacy options.
The Backblaze-Vultr partnership means more developers can build the flexible tech stacks theywant to build, without having to make needlessly tough choices between access andaffordability, said Shane Zide, VP of Global Partnerships at Vultr. When two companies whofocus on ease-of-use and price-performance make their technologies work together, the wholeis greater than the sum of the parts.
In todays landscape for data storage, having more options is always a fantastic idea. Vultr is the largest privately owned global hyper-scale cloud provider on the market, so it becomes a nice alternative to feeling forced into using Amazon, Google, or Microsoft cloud business storage for scalability.
Vultr has more than 1.3 million customers, with more than 40 million instances deployed across 17 global environments. Vultr brings an affordable and easy-to-use cloud environment for businesses looking to streamline operations and avoid working with the biggest tech giants on the market.
For more information, check out the Backblaze blog.
Check out 9to5Mac on YouTube for more Apple news:
Public cloud hits the spot for NetApp’s Q122 results Blocks and Files – Blocks and Files
NetApp grew its revenues 12 per cent year-on-year to $1.46 billion in its first fiscal 2022 quarter, ended July 30, and near the high end of its guidance, with profits up 162.3 per cent to $202 million.
This was not as high a growth rate as Pures 23 per cent, but far better than Dells one per cent decline in its storage revenues.
The announcement statement by CEO George Kurian was straightforward: Building on our accelerating momentum through last year, we are off to a great start to fiscal 2022, with strong revenue, gross margin, and operating leverage across the entire business.
He was much more enthusiastic in the earnings call: We just finished a phenomenal quarter. In Q1, we grew our all-flash business by 23 per cent, overall product revenue by 16 per cent and cloud by 155 per cent year-on-year. We raised our fiscal year to eight to nine per cent growth and anticipate delivering close to $5 in earnings per share. These are all record numbers, right, operating margins, full year and earnings per share for the company. So, I feel very, very, very good about where we are.
Kurian is confident that NetApp gained share in the storage and all-flash markets.
Financial summary:
A chart shows the renewed growth pattern in its quarterly revenues, with five successive growth quarters since its fiscal 2020 year commenced:
There is still some way to go to get back to fiscal 2012 and 2013 revenue levels, but we are seeing a strong recovery as the COVID pandemic plays out.
NetApp saw strength and momentum with All-Flash Arrays (AFA), which now have a $2.8 billion run rate although that was down from last quarters $2.9 billion. NetApps AFAs have a 29 per cent penetration of the installed ONTAP base, so there is still a lot of latent demand ahead.
CFO Mike Berry said We ended Q1 with over $3.9 billion in deferred revenue an increase of eight per cent year-over-year. Q1 marks the 14th consecutive quarter of year-over-year deferred revenue growth, which continues to be the best indicator of the health of our recurring revenue.
NetApp rejigged its reporting segments with two new classifications: Hybrid Cloud and Public Cloud. The Hybrid Cloud segment covers on-premises/private cloud storage and data management products, such as ONTAP and StorageGRID arrays, with links to the public cloud. The Public Cloud includes products delivered as-a-service from public clouds, such as Spot.
Hybrid Cloud revenues in the quarter were $1.38 billion up 8.7 per cent on a year ago and driven by AFA growth. The much smaller Public Cloud segment pulled in just $79 million, up a huge 154.8 per cent over the same period. This was driven by driven by Cloud Volumes, Cloud Insights, and Spot by NetApp services. Public Cloud annual recurring revenue rose 89 per cent to $337 million.
We think that NetApp must have high hopes indeed for its Public Cloud segment to be be separating it out at such a low revenue point. It is a completely different business from its traditional storage business as we pointed out in July when NetApp added a Windows 365/CloudPC offering to its Spot cloud broking business.
Kurian said Our Public Cloud services not only allow us to participate in the rapidly growing cloud market, they also make us a more strategic data center partner to our enterprise customers, driving share gains in our Hybrid Cloud business.
He is confident that the Public Cloud business will grow. I think we see that over the next few years, as businesses deploy more of their core operations on public clouds, the opportunities we have around compute automation management through Spot, storage and data protection and management through Cloud Volumes and monitoring through Cloud Insights will be a very, very strong business. And all of the hyperscaler partners that we work with see it much the same way, which is why theyre building more and more capabilities with us.
We also see a new growth engine in the Public Cloud segment with all of the cloud native work we are doing around containers and serverless.
The Public Cloud focus is influencing M&A, with Berry pointing out: Our acquisition strategy will remain focused on bolstering our strategic Public Cloud roadmap.
The outlook for the second FY22 quarter is $1.54 billion +/- $50 million an 8.8 per cent rise on a year ago. The full FY22 outlook is for revenues to be eight to nine per cent higher than FY21s $5.74 billion, meaning $6.2 billion to $6.26 billion.
See the rest here:
Public cloud hits the spot for NetApp's Q122 results Blocks and Files - Blocks and Files
HUAWEI AppGallery Connect Serverless Services Officially Released to Achieve Easy App Development and O&M – PRNewswire
Auth Service enables developers to quickly build a user authentication system at lower costs by providing a pre-built, hosted authentication system. It provides SDKs for different platforms, allowing users to sign into apps using a cellphone number, email address, HUAWEI ID, or mainstream third-party accounts such as Facebook and Twitter for a smooth sign-in experience.
Cloud Functions helps to quickly build app backend servicesallowing code to be run without having to manage any servers. With it, development and testing are carried out by function, so developers can focus on service logic development. It also helps to simplify O&M and allocate resources properly.
Cloud DB, a scalable serverless database, offers secure and reliable data management services. It provides easy-to-use cloud and device SDKs.And it ensures data can be automatically synchronized among different devices, meaning you can use Cloud DB to quickly develop secure and reliable apps.
Cloud Storage allows developers to securely and economically store high volumes of data such as images, audio, and video generated by users.It provides cloud and device SDKs so developers do not need to set up servers or make content delivery network (CDN) configurations, and O&M can be completed automatically.
Cloud Hosting provides one-stop hosting capabilities enabling quick website release. That is, users can access web apps without needing to deploy any cloud servers. With Cloud Hosting, developers can focus only on building service logic instead of needing to pay attention to deployment details over the cloud.
In conclusion, HUAWEI AppGallery Connect Serverless Services offers:
1. Pay-as-you-go mode
2. O&M-free services
3. Fast rollout capability
We are now offering free quotas for developers to try out these services. For more details, please refer to the following documents: Auth Service, Cloud Functions, Cloud Storage, Cloud DB, and Cloud Hosting. You can also explore more on Github.For any questions, please send an email to [emailprotected].
Photo - https://mma.prnewswire.com/media/1607477/AppGallery_Connect_Serverless_Services.jpg
SOURCE Huawei consumer business group
Go here to see the original:
HUAWEI AppGallery Connect Serverless Services Officially Released to Achieve Easy App Development and O&M - PRNewswire
Microsoft Teams: Here are the features added in the past month – ZDNet
Microsoft continues to bolster Teams features for organizations with hybrid or remote-work policies, including better live transcriptions, Android features, and smarts for the education sector.
Microsoft's latest Teams updates as detailed in its August update should make video meetings more palatable for everyone and includes new features for presenters and participants. Live captions and live transcriptions in Teams are handy for multilingual global firms and anyone who needs to take notes during meetings.
The best cloud storage services
Free and cheap personal and small business cloud storage services are everywhere. But, which one is best for you? Let's look at the top cloud storage options.
Read More
Teams now has "reporter mode" that positions content on a screen that's shared with participants, such as a host's slides, as a visual aid above the presenter's shoulder, just like a news story would be presented. There's also a side-by-side mode that displays a video feed next to the presenter's content.
SEE: When the return to the office happens, don't leave remote workers out in the cold
These two new presenter modes join the previously announced Standout Mode, and are available when sharing a desktop and window in the Teams desktop app.
Microsoft launched its AI-powered live transcription service in Teams in March for Microsoft 365 E3, Microsoft 365 E5, Microsoft 365 Business Standard, and Microsoft 365 Business Premium SKUs. Rival video-meeting service Zoom brought live transcriptions to free accounts.
But Microsoft's Teams real-time captioning and transcription feature was previously only available in English. Now it's also available in 27 more languages, including German, Portuguese, Japanese, and Hindi.
Another live transcript feature that has come to Teams is that the app will immediately start live transcription when recording is turned on. Admins need to enable Allow Transcription and Allow cloud-recording policies, then the transcription can accessed through the Recordings & Transcripts tab of a meeting.
There are also updates for Teams Rooms on Android in version 1449/1.0.96.2021070803, which allows for faster login and more structured deployment of Android devices. It means admins can remotely provision and sign-in to Teams panels.
Teams Together Mode can now be customized through Microsoft's low-code Power Platform. It's one more way Microsoft is integrating Teams, Outlook and Dynamics 365.
SEE: Windows 11 FAQ: Here's everything you need to know
As ZDNet's Mary Jo Foley reported in August, Microsoft will also slash its Power Apps per-user prices by half on October 1 from from $10 per month to $5 per month.
Other updates include the Teams for Education 'reading progress' tool, which it launched in August. It's aimed at helping teachers to assess the reading fluency of students in video and audio formats.
The rest is here:
Microsoft Teams: Here are the features added in the past month - ZDNet
Quobyte: AI supports sustainable logging – saving owls’ trees from the axe Blocks and Files – Blocks and Files
Case Study. Northern Spotted Owls, a federally protected species, live in the Oregon forests forests where loggers want to cut down trees and harvest the timber. The areas where the owls live are protected from the loggers, but which areas are they? Data can show the way, enabling sustainable timber harvesting.
Recorded wild life sounds, such as birdsong, are stored with Quobyte file, block and object software and interpreted using AI. This data shows where the owl habitats are located
Forests of Douglas Fir, Ponderosa Pine, Juniper and other mixed conifers cover over30.5 million acres of Oregon almost half of the state. Finding out where the owls live is quite a task. The Center for Quantitative Life Sciences (CQLS) at Oregon State University, working with the USDA Forestry Service, has deployed and is tracking 1500 autonomous sound and video recording units deployed in the forests, gathering real-time data. The aim is to to monitor the behaviour of wildlife living in the forests of Oregon to ensure that the logging industrys impact is managed.
The CQLS creates around 250 terabytes of audio and video data a month from the recording units, and maintains around 18PB of data at any given time. It keeps taking data off and reusing the space to avoid buyinginfinite storage.
Over an 18-month period it devised an algorithm to parse the audio recordings and identify different animal species. The algorithm creates spectrograms from the audio, and processes those spectrograms through a convolutional neural net based on the video. It can identify about thirty separate species, distinguish male from female, and even spot behavioural changes within a species over time.
The compute work takes place on an HPC setup comprising IBM Power System AC922 servers, collectively containing more than 6000 processors across 20 racks in two server rooms that serve 2500 users. The AC922 architecture puts AI-optimised GPU resources directly on the northbridge bus, much closer to the CPU than conventional server architectures.
CQLS needed a file system and storage solution able to keep massive datasets close to compute resources as swapping data in and out from external scratch resources doubled processing time.
At first it was looking at public cloud storage options, but the costs associated were considered outrageously expensive.
CQLS checked a variety of storage alternatives and settled on Quobyte running on COTS hardware, rejecting more expensive storage appliance alternatives which could need expensive support arrangements.
The sizes of individual files vary from tiny to very large and everything in between. The Quobyte software is good when dealing with large files, as opposed to millions of highly similar small files. This is advantageous when working on AI training, where TIFF files can range from 20 to 200GB in size.
Concurrently, those files may need to be correlated with data from sensors, secondary cameras, microphones, and other instruments. Everything must flow through one server, which puts massive loading on compute and storage.
Quobytes software uses four Supermicro servers with two Intel Xeon E5-2637 v4 CPUs @ 3.50GHz and 256G RAM (DDR4 2400). There are LSI SAS3616 12Gbit/s SAS controllerd running two 78-disk JBODs. These are filled with Toshiba MG07ACA14TA 14TB, SATA-6Gbit/s, 7200rpm, 3.5-inch disk drives.
The entire HPC system is Linux-based and everything is mountedthrough the Quobyte client for x86-based machines and NFS for the PPC64LE (AC922) servers.
Many groups of users access the system. A single group could have millions or hundreds of files based on the work they do. Most groups leverage over 50TB each and currently there is 2.6PB loaded on the Quobyte setup.
Christopher Sullivan, Assistant Director for Biocomputing at CQLS, said; We have all kinds of pathways for data to come into the systems. First off all research buildings at OSU are connected at a minimum of 40Gbit/sec network and our building and incoming feed to the CGRB (Center for Genome Research and Biocomputing) is 100Gbit/sec and a 200Gbit/sec network like between OSU and HMSC (Hatfield Marine Science Center) at the coast.
To start some of the machines in our core lab facility (not the sequencers) do drop live data onto the system through SAMBA or NFS-mounted pathways. Next, we have users moving data onto the servers via a front-end machine, again providing services like SAMBA and SSH with a 40Gbit/sec network connection for data movement.
This allows for users to have machines around the university move data automatically or by hand onto the systems. For example, we have other research labs moving data from machines or data collected in greenhouses and other sources. The data line to the coast mentioned above is used to move data onto the Quobyte for the plankton group as another example.
What about backup?
Sullivan said: Backup is something we need on a limited basis since we can generally generate the data again cheaper than the costs of paying for backing up that large amount of space. Most groups backup the scripts and final output (ten per cent) of the space they use for work. Some groups take the original data and if needed by grants keep the original data on cold drives on shelves in a building a quarter-mile away from the primary. So again we do not need a ton of live backup space.
Sullivan said: We found that using public clouds was too expensive since we are not able to get the types of hardware in spot instances and data costs are crazy expensive. Finally, researchers cannot always tell what is going to happen with their work or how long it needs to run, etc.
This makes the cloud very costly and on-premises very cost-effective. My groups buy big machines (256 thread count with 2TB RAM and 12x GPUs) that last 67 years and can do anything they need. That would be paid for five times over in that same time frame in the cloud for that hardware. Finally, the file space is too expensive over the long haul, and hard to move large amounts of data on and off. We have the Quobyte to reduce our overall file space costs.
This is a complicated and sizeable HPC setup which does more than safeguard the Northern Spotted Owls arboreal habitats. That work is done in an ingenious way one that doesnt involve human bird-spotters trekking through the forests looking for the creatures.
Instead, AI algorithms analyse and interpret recorded bird songs, recognise those made by the owls and then log where and when the owls are located. That data can be used to safeguard areas of forests from the loggers, who can fell timber sustainably, cutting down owl-free timber.
The rest is here:
Quobyte: AI supports sustainable logging - saving owls' trees from the axe Blocks and Files - Blocks and Files