Earlier in January, StorPool announced the new capabilities added in the scope of StorPool Storage 19.2 and our shift to a process of rapid continuous improvement of the most sophisticated block storage platform om the market.
In this article, we cover some of the recent updates to the existing capabilities of StorPool Storage platform. Since we introduced many enhancements last year, this article is the first in a series of five pieces.
Today, our focus is on the three core capabilities of a storage platform serving block storage to primary and secondary user workloads in a public or private cloud environment:
Reliability – you need a data storage platform that delivers the peace of mind that your cloud infrastructure’s foundation is robust. Storage is the infrastructure piece all end-user workloads rely on to operate consistently. Even for secondary storage systems, you need certainty that you can access your data if you have a disaster scenario in your production environment. Therefore, the storage systems deployed using a given data storage solution must keep receiving and serving data with end-to-end integrity – irrespective of factors like software updates, hardware bugs or failures, firmware, driver, or kernel issues, server or network upgrades, and so on. Even in cases beyond anyone’s control, such as catastrophic power failure or complete network downtimes, after order is restored, users must be able to carry on from the last stable state.
Agility – you need a data storage platform that grows with your business and enables you to address user requirements online, without disrupting their workloads. Therefore, your storage solution should allow you to convert sets of standard servers, with components matching your use case requirements, into primary or secondary storage systems. You must be able to adjust the scale, performance, and resource consumption of your storage system in-flight – by adding drives or servers, by changing workload settings like IOPS, virtual disk size, and data placement policies, or by serving storage to multiple hypervisors and Cloud Management Platforms from a single storage system. To guarantee that there is no vendor lock-in, all of this and more must be enabled using software that manages your servers’ standard components.
Speed – your data storage platform must deliver the performance needed to meet your use cases’ demands. It should not cause bottlenecks, even when you are using only mass-produced, industry-standard components. Your primary and secondary storage systems’ speed is critical if you want to accelerate your services, deliver a snappy experience to your customers, and delight the users of your applications. Therefore, your storage solution must be capable of extracting the maximum possible speed (latency, IOPS, data throughput) for a given set of industry-standard servers. At StorPool, we have designed and developed unique technologies to ensure that we aggregate your storage drives’ performance linearly. The result – as a given StorPool-powered storage system scales, IOPS and throughput increase for all the data inside while latency remains low (~100 µs for NVMe SSDs, sub-200 µs for SATA SSDs).
Since all the pieces of modern cloud infrastructure rely on the storage system, we believe that there is no room for trade-offs between the reliability, agility, and speed of a data storage solution. That is why we have invested in building a solution with a compelling mesh of these three capabilities. We continue to raise our standards and invest in building cutting-edge features that guarantee our customers’ primary and secondary storage systems are running optimally.
The result of all this – StorPool simplifies our customers’ cloud infrastructure, removes all the pains they experienced with legacy storage products, and supercharges their revenue and profits.
Read on to get an overview of the recent improvements to the reliability, agility, and speed of StorPool Storage.
- Creating Volumes is now Allowed even in a Critical System State – Enables users to create volumes and serve them to virtual machines, even in a situation where one node in a three-node cluster has failed (e.g., motherboard failure). While the system is in a critical state, StorPool keeps two of the data replicas on one of the two remaining nodes. When the missing node returns online, the copies are re-distributed to all nodes.
- Improved quorum stability in unstable network conditions.
- Fixed an issue in automatic detach after rebalancing completes for soft-ejected drives.
- Improved storage capacity and performance aggregation logic during drive and node recovery.
- Added Option for Silent Initialization of Storage Drives – the –no-notify option makes it easier to change attributes of drives added to StorPool.
- Changes to Multi-site and Multi-cluster – added improvements and stability fixes based on insights gained from several multi-petabyte deployments – made volume create/update flags mandatory for multi-cluster API endpoints, added function for reverting volumes to snapshots in a multi-cluster deployment, added ability to list exported volumes and remote volumes, and more.
- Added additional validations in the initidisk tool used for storage drive initialization and various other improvements of initdisk tool.
- Prepared StorPool Storage API for the introduction of multi-cluster support in Cloud Management Platform integrations.
- Improved Sustained Write Performance – improved aggregation performance and efficiency, which permits processing more end-user write operations.
- Improved performance under mixed production loads.
- Improved API responsiveness during dematerialization
- Fixed balancer behavior for cases where overrides are enabled because many volumes were created from the same template.
- Improved performance when relocating data from SATA or NVMe SSDs in case of rebuild or rebalancing of data.
- Improved the efficiency of the bridge connection and cross-cluster mgmt service interactions in case of multi-site deployments.
Conclusion – What a Modern Data Storage Platform Should Do
The International Data Corporation estimates that, by 2024, 60% of annual IT infrastructure spend will be for private cloud and public cloud computing models. Deploying a cloud requires sets of hardware and software pieces, commonly known as cloud infrastructure – servers, storage, network switches, virtualization software, service tools, and management tools like Cloud Management Platforms (e.g., OpenNebula, OnApp, CloudStack, OpenStack).
All the pieces in a public or private cloud deployment connect to the storage and rely on its performance to operate consistently. StorPool Storage is a solution that excels in the three core capabilities of data storage – reliability, agility, and speed. It is an ideal foundation for cloud infrastructure running diverse, mission-critical workloads.
In the next article of this series, we will talk about the improvements in the Management and Monitoring capabilities of StorPool Storage. Check-in with us soon or subscribe to our mailing list if you want to get notified about future articles.