Announcing the general availability of Azure Ultra Disk Storage

Today, we are announcing the general availability (GA) of Microsoft Azure Ultra Disk Storage—a new Managed Disks offering that delivers unprecedented and extremely scalable performance with sub-millisecond latency for the most demanding Azure Virtual Machines and container workloads. With Ultra Disk Storage, customers are now able to lift-and-shift mission critical enterprise applications to the cloud including applications like SAP HANA, top tier SQL databases such as SQL Server, Oracle DB, MySQL, and PostgreSQL, as well as NoSQL databases such as MongoDB and Cassandra.With the introduction of Ultra Disk Storage, Azure now offers four types of persistent disks—Ultra Disk Storage, Premium SSD, Standard SSD, and Standard HDD. This portfolio gives our customers a comprehensive set of disk offerings for every workload.

Ultra Disk Storage is designed to provide customers with extreme flexibility when choosing the right performance characteristics for their workloads. Customers can now have granular control on the size, IOPS, and bandwidth of Ultra Disk Storage to meet their specific performance requirements. Organizations can achieve the maximum I/O limit of a virtual machine (VM) with Ultra Disk Storage without having to stripe multiple disks. Check out the blog post “Azure Ultra Disk Storage: Microsoft's service for your most I/O demanding workloads” from Azure’s Chief Technology Officer, Mark Russinovich, for a deep under-the-hood view.

Since we launched the preview for Ultra Disk Storage last September, our customers have used this capability on Azure on a wide range of workloads and have achieved new levels of performance and scale on the public cloud to maximize their virtual machine performance.

Below are some quotes from customers in our preview program:

“Ultra Disk Storage enabled SEGA to seamlessly migrate from our on-premise datacenter to Azure and take advantage of flexible performance controls.”

– Takaya Segawa, General Manager/Creative Officer, SEGA

“Ultra Disk Storage allows us to achieve incredible write performance for our most demanding PostgreSQL database workloads – giving us the ability to scale our applications in Azure.” 

– Andrew Tindula, Senior IT Manager, Online Trading Academy

Ultra Disk Storage performance characteristics

Ultra Disk Storage offers sizes ranging from 4 GiB up to 64 TiB with granular increments. In addition, it is possible to dynamically configure and scale the IOPS and bandwidth on the disk independent of capacity.

Customers can now maximize disk performance by leveraging:

Up to 300 IOPS per GiB, to a maximum of 160K IOPS per disk
Up to a maximum of 2000 MBps per disk

Pricing and availability

Ultra Disk is now available in East US 2, North Europe, and Southeast Asia. Please refer to the FAQ for latest supported regions. For pricing details for Ultra Disk, please refer to the pricing page. The general availability price takes effect from October 1, 2019 unless otherwise noted. Customers in preview will automatically transition to GA pricing on this date. No additional action is required by customers in preview.

Get started with Azure Ultra Disk Storage

You can request onboarding to Azure Ultra Disk Storage by submitting an online request or by reaching out to your Microsoft representative.
Quelle: Azure

Azure Ultra Disk Storage: Microsoft's service for your most I/O demanding workloads

Today, Tad Brockway, Corporate Vice President, Microsoft Azure, announced the general availability of Azure Ultra Disk Storage, an Azure Managed Disks offering that provides massive throughput with sub-millisecond latency for your most I/O demanding workloads. With the introduction of Ultra Disk Storage, Azure includes four types of persistent disk—Ultra Disk Storage, Premium SSD, Standard SSD, and Standard HDD. This portfolio gives you price and performance options tailored to meet the requirements of every workload. Ultra Disk Storage delivers consistent performance and low latency for I/O intensive workloads like SAP Hana, OLTP databases, NoSQL, and other transaction-heavy workloads. Further, you can reach maximum virtual machine (VM) I/O limits with a single Ultra disk, without having to stripe multiple disks.

Durability of data is essential to business-critical enterprise workloads. To ensure we keep our durability promise, we built Ultra Disk Storage on our existing locally redundant storage (LRS) technology, which stores three copies of data within the same availability zone. Any application that writes to storage will receive an acknowledgement only after it has been durably replicated to our LRS system.

Below is a clip from a presentation I delivered at Microsoft Ignite demonstrating the leading performance of Ultra Disk Storage:

Microsoft Ignite 2018: Azure Ultra Disk Storage demo

Below are some quotes from customers in our preview program:

“With Ultra Disk Storage, we achieved consistent sub-millisecond latency at high IOPS and throughput levels on a wide range of disk sizes. Ultra Disk Storage also allows us to fine tune performance characteristics based on the workload.”

– Amit Patolia, Storage Engineer, DEVON ENERGY

“Ultra Disk Storage provides powerful configuration options that can leverage the full throughput of a VM SKU. The ability to control IOPS and MBps is remarkable.”

– Edward Pantaleone, IT Administrator, Tricore HCM

Inside Ultra Disk Storage

Ultra Disk Storage is our next generation distributed block storage service that provides disk semantics for Azure IaaS VMs and containers. We designed Ultra Disk Storage with the goal of providing consistent performance at high IOPS without compromising our durability promise. Hence, every write operation replicates to the storage in three different racks (fault domains) before being acknowledged to the client. Compared to Azure Premium Storage, Ultra Disk Storage provides its extreme performance without relying on Azure Blob storage cache, our on-server SSD-based cache, and hence it only supports un-cached reads and writes. We also introduced a new simplified client on the compute host that we call virtual disk client (VDC). VDC has full knowledge of virtual disk metadata mappings to disks in the Ultra Disk Storage cluster backing them. That enables the client to talk directly to storage servers, bypassing load balancers and front-end servers used for initial disk connections. This simplified approach minimizes the layers that a read or write operation traverses, reducing latency and delivering performance comparable to enterprise flash disk arrays.

Below is a figure comparing the different layers an operation traverses when issued on an Ultra disk compared to a Premium SSD disk. The operation flows from the client to Hyper-V to the corresponding driver. For an operation done on a Premium SSD disk, the operation will flow from the Azure Blob storage cache driver to the load balancers, front end servers, partition servers then down to the stream layer servers as documented in this paper. For an operation done on an Ultra disk, the operation will flow directly from the virtual disk client to the corresponding storage servers.

Comparison between the IO flow for Ultra Disk Storage versus Premium SSD Storage

One key benefit of Ultra Disk Storage is that you can dynamically tune disk performance without detaching your disk or restarting your virtual machines. Thus, you can scale performance along with your workload. When you adjust either IOPS or throughput, the new performance settings take effect in less than an hour.

Azure implements two levels of throttles that can cap disk performance, a “leaky bucket” VM level throttle that is specific to each VM size, described in documentation. A key benefit of Ultra Disk Storage is a new time-based disk level throttle that is applied at the disk level. This new throttle system provides more realistic behavior of a disk for a given IOPS and throughput. Hitting a leaky bucket throttle can cause erratic performance, while the new time-based throttle provides consistent performance even at the throttle limit. To take advantage of this smoother performance, set your disk throttles slightly less than your VM throttle. We will publish another blog post in the future describing more details about our new throttle system.

Available regions

Currently, Ultra Disk Storage is available in the following regions:

East US 2
North Europe
Southeast Asia

We will expand the service to more regions soon. Please refer to the FAQ for the latest on supported regions.

Virtual machine sizes

Ultra Disk Storage is supported on DSv3 and ESv3 virtual machine types. Additional virtual machine types will be supported soon. Refer to the FAQ for the latest on supported VM sizes.

Get started today

You can request onboarding to Azure Ultra Disk Storage by submitting an online request or by reaching out to your Microsoft representative. For general availability limitations refer to the documentation.
Quelle: Azure

AWS Storage Gateway unterstützt jetzt Tag-on-Create und die Tag-basierte Zugriffskontrolle

AWS Storage Gateway unterstützt jetzt das Hinzufügen von Tags zum Zeitpunkt der Ressourcenerstellung sowie die Tag-basierte Zugriffskontrolle. Sie können allen Storage Gateway-Ressourcen beim Erstellen Tags zuweisen – einschließlich Gateways, Dateifreigaben, Volumes, Volume-Snapshots und Bändern. Zudem können Sie differenzierte Zugriffssteuerungen definieren, indem Sie Tags und IAM-Richtlinien (AWS Identity and Access Management) verwenden, um den Zugriff auf Ressourcen und Aktionen zu steuern. Tags sind einfache Schlüsselwertpaare, die Sie Ressourcen zuweisen können, um Ressourcen einfach zu organisieren, zu suchen und zu identifizieren, Kostenzuordnungsberichte zu erstellen und den Zugriff auf Ressourcen zu steuern.
Quelle: aws.amazon.com

Wir stellen vor: Der neue AWS Co-Branding-Leitfaden

Wir stellen den neuen AWS Co-Branding-Leitfaden vor, der APN-Partnern Richtlinien für das gemeinschaftliche Marketing mit AWS bietet. Dieser Leitfaden enthält bewährte Methoden, um Ihre Marke in genehmigten gemeinschaftlichen Marketing-Kampagnen effektiv neben AWS zu präsentieren. Im AWS Co-Branding-Leitfaden erfahren Sie unter anderem verschiedene Möglichkeiten des Co-Brandings mit AWS, die Anforderungen für ein AWS Logo Lockup, Design- und Nutzungsrichtlinien, wichtige Hinweise zum Co-Branding und erhalten eine Co-Branding-Checkliste. Der AWS Co-Branding-Leitfaden richtet sich an APN-Partner der Stufe „Select“ oder höher mit genehmigten Co-Branding-Aktivitäten. Informationen zu genehmigten Co-Branding-Aktivitäten und zur Nutzung von Co-Branding-Materialien erhalten Sie hier im Leitfaden. Wir empfehlen zudem die Teilnahme an einer Schulung am 15. August, um mehr über den neuen AWS Co-Branding-Leitfaden zu erfahren. Auch im APN Blog finden Sie detaillierte Informationen zum neuen Leitfaden.
Quelle: aws.amazon.com