StorPool Storage – the leading software-defined storage provider for private cloud and public cloud builders is pleased to announce StorPool’s new release – 18.02. The new version of the fast distributed block storage incorporates industry-leading capabilities into the solution providing everything needed to ensure efficient, reliable and scalable storage solution for your business.
The distributed storage software, developed by StorPool is recognized by Service Providers and Enterprises as one of the fastest and most reliable solutions in the domain, due to its low latency and sophisticated data redundancy features. With the new release, StorPool puts an emphasis on new integrations, extended hardware compatibility, hardware acceleration, and reduced resource usage.
“StorPool relies on its experience with hundreds of public and private clouds. We added optimal use of hyperscale leaf-spine networks, an integration into the leading container orchestration platform Kubernetes, and a myriad of small improvements”, said Boyan Krosnov, CPO of StorPool. “StorPool’s performance was exceptional, but we like to push the boundaries of what’s possible. That’s why we implemented what we call “accelerated mode” for NICs. We can now complete storage operations in microseconds of latency, without sacrificing any of the advanced features of our solution, which rival a high-end SAN. It is not about “hero IOPS numbers”. Consistent low storage latency makes applications run up to 8 times faster.”
What you will find in StorPool’s New Release:
StorPool provides reliable persistent block-storage for Kubernetes. Kubernetes is an open-source project that came out of Google – a container orchestrator which has a built-in support for running and managing containers, their monitoring, load balancing, and auto-scaling.
Kubernetes is a way to run modern cloud-native applications in large clusters by abstracting the underlying hardware and other infrastructure. It provides resource scheduling, horizontal scaling, life-cycle management and many more.
The integration enables Kubernetes to create and use the storage volumes provided by a StorPool storage cluster and to seamlessly migrate the workloads between container servers.
Enhanced hardware platform compatibility and ARM and POWER support
We extended our hardware support list. In this way, we enable our customers to build even more sophisticated infrastructure and storage systems on the latest generation of hardware. Some of the newly supported hardware components are:
- 4Kn drives – more information about them can be found here
- Hot-plug NVMe – making it easier to manage the infrastructure without rebooting servers
- ARM and POWER 8/9 architectures
Traditionally, ARM is associated with mobile phones because of its low power consumption and its efficiency. Recently, the IT infrastructure industry recognizes ARM server processors as a challenger to the x86 server architecture. It may become common in public cloud, IoT, telco, and edge computing.
The POWER processor architecture was originally developed for enterprise workloads. the customers are banks, telcos, and companies with a demand for powerful servers to build complex infrastructure. The POWER architecture is gaining wider interest because POWER CPUs are now being integrated into third-party servers through the OpenPOWER initiative.
30% lower storage latency
For the majority of workloads and applications, latency is the most important metric. It is more important than IOPS. Since many applications are performing inter-dependent storage operations – the lower the latency of the storage system, the faster the applications are.
StorPool 18.02 release has some 30% lower storage operations latency, compared to the previous version. We’ve achieved this by developing an accelerated mode (also known in the industry as kernel bypass). The accelerated mode is supported for Intel, Mellanox, and Broadcom NICs.
With accelerated mode, RDMA and NVMe drives, a StorPool storage system, with triple redundancy across servers, end-to-end data integrity, and advanced data management features commits 4K write operations in approximately 50 microseconds (0.05 ms). This is something really unique on the market.
New Linux distributions supported
The new release of StorPool is coming with support for new distributions. The major one is Ubuntu 18.04 LTS. As always, we are keeping up with updates to major supported distributions to make sure we offer the best storage solution for our customers.
Support for hyperscale datacenter networks
Unlike smaller public and private clouds, hyperscale deployments typically use a routed leaf/spine datacenter networks.
In this release, StorPool improved its support of routed leaf/spine datacenter networks, so to address the raising demand for hyperscale deployments. This is in addition to the simpler and proven flat Ethernet topology, which is used in small to mid-sized deployments.
Enhanced analytics and metrics collection
Having full insight into your storage and infrastructure is critical for detailed analysis. This enables data-driven decisions and leveraging data to increase efficiency and cut costs. StorPool’s new release, 18.02 includes even better analytics with detailed per-server CPU and memory statistics. The CPU analytics feature enables our customers to monitor CPU load and congestion on hypervisor nodes. While CPU on hypervisors is not strictly a storage system metric, we’ve found that it is a common cause for complaints about “slower storage”, hence we’ve extended metrics collection to cover it too.
Many other improvements have been done to fine-tune your storage experience:
- reliability and performance improvements of the multi-site connectivity
- Lower CPU usage and higher efficiency for storpool_block during TRIM operations
- A new tool added – transfer_status to show the status of the currently running bridge transfers
- A new tool added – sDump to print network stats in a human-readable format
- StorPool CLI prompt available to be customized using the SP_CLI_PROMPT variable
- Improved fio test scripts
- A new tool, storpool_ping, for measuring latency and throughput across the storage network, and a higher level tool for quickly checking network connectivity and performance between all storage server and initiator nodes
For more questions, do contacts us at [email protected]