Anton Kolomyeytsev
Anton Kolomyeytsev
Anton Kolomyeytsev is StarWind CTO, Chief Architect & Co-Founder. Microsoft Most Valuable Professional [MVP] in Cluster 2014 & 2015. SMB3, NFS, iSCSI & iSER, NVMe over Fabrics.

All posts by this author

Posted by Anton Kolomyeytsev on June 23, 2016
ReFS: Performance

ReFS (Resilient File System – is a Microsoft file system, which ensures data integrity by means of resiliency to corruption (irrespective of software or hardware failures), increases data availability and scales to large data sets across various workloads. Its data protection feature is represented by the FileIntegrity option, which is responsible for file scanning and repair processes.

Learn More

Posted by Anton Kolomyeytsev on June 16, 2016
Software-Defined Storage: StarWind Virtual SAN vs Microsoft Storage Spaces Direct vs VMware Virtual SAN

This is a comprehensive comparison of the leading products of the Software-Defined Storage market, featuring Microsoft Storage Spaces Direct, VMware Virtual SAN and StarWind Virtual SAN. It provides numerous use cases, based on different deployment scales and architectures, because the mentioned products all have different aims. As the market is already large enough, the vendors used to dwell its different parts, but lately they entered a full-scale competition, adapting their products to meet general demand. This post is an analysis of how Microsoft, VMware and StarWind fare in in the Software-Defined Storage market right now. The approach is practical and all the statements are based on the experience of virtualization administrators and engineers from all over the world.


Learn More

Posted by Anton Kolomyeytsev on May 11, 2016
Storage Replica: Overview

Here is an overview dedicated to disaster recovery, more specific, it’s about the DR capabilities of Microsoft Storage Replica – a new feature of Windows Server 2016. It takes a glance on the DR process itself and then brings a few details of the Storage Replica operation, its features, and peculiarities. They include zero data loss, block-level replication, simple deployment and management, guest and host, SMB3 protocol, high security, high performance, consistency groups, user delegation, network constraint, thin provisioning, etc. The post is, basically, an introduction to a series of experiments also listed on the blog. They were conducted in order to check the functionality and performance of Microsoft Storage Replica in different use cases.

Asynchronous and Synchronous replication

Learn More

Posted by Anton Kolomyeytsev on May 10, 2016
SMB3: Overview

This is an overview of the Server Message Block (SMB3) protocol from Microsoft. It offers a short insight into the history of SMB3 creation and development over the years (as the idea is technically around 30 years old). As of Windows Server 2012, the protocol got new features: SMB Transparent Failover, SMB Scale Out, SMB Multichannel, SMB Direct, SMB Encryption, VSS for SMB file shares, SMB Directory Leasing, SMB PowerShell. In Windows Server 2016, it also got Pre-authentication integrity and Cluster dialect fencing. The post concentrates on RDMA-capable SMB Direct and MPIO-utilizing SMB Multichannel and their benefits. Also, it is an introduction to a series of tests aimed at creating SMB 3.0 File Servers in an unusual way.


Learn More

Posted by Anton Kolomyeytsev on April 12, 2016
ReFS: Log-Structured

Here is a part of a series about Microsoft Resilient File System, first introduced in Windows Server 2012. It shows an experiment, conducted by StarWind engineers, dedicated to seeing the ReFS in action. This part is mostly about the FileIntegrity feature in the file system, its theoretical application and practical performance under real virtualization workload. The feature is responsible for data protection in ReFS, basically the reason for “resilient” in its name. It’s goal is avoidance of the common errors that typically lead to data loss. Theoretically, ReFS can detect and correct any data corruption without disturbing the user or disrupting production process.

Device manager

Learn More

Posted by Anton Kolomyeytsev on April 9, 2016
ReFS: Overview

This is a short overview of Microsoft Resilient File System, or ReFS. It introduces the subject and gives a short insight into its main characteristics and theoretical use. It is a part of a series of posts dedicated to ReFS and is, basically, an introduction to the practical posts. All the experiments that show how ReFS really performs, are also listed in the blog. ReFS seems to be a great replacement for the NTFS and its resilience is most convenient for cases, when data loss is critically unacceptable. The file system cooperates with Microsoft Storage Spaces Direct in order to perform automatic corruption repairs, without any attention of the user.


Learn More

Posted by Anton Kolomyeytsev on January 22, 2016
Hyper-V: Free “Shared Nothing” SMB3 Failover File Server

A part of a series, the research is dedicated to the capability of free Microsoft Hyper-V Server R2 to assume a file server role and cluster the resulting file server. Our last experiment on this matter showed that it’s possible to create SMB3 File Server on the free Microsoft Hyper-V Server R2 and now we’re about to go further and cluster it. The post shows a detailed instruction on the process and also the resulting setup. It appears to be fully working and usable, so the process may be called a success after all. Though it is absolutely real to do so, it is a violation of Microsoft’s license agreement, so StarWind urges everyone to refrain from repeating the experiment.

2 servers with Microsoft Hyper-V Server 2012 R2 added into domain and the client node with Windows Server 2012 R2

Learn More

Posted by Anton Kolomyeytsev on January 22, 2016
Hyper-V: Free SMB3 File Server

The first experiment dedicated to building a file server on free Microsoft Hyper-V Server R2. It is a part of a series of similar practical posts. The post offers a detailed instruction on how to assign file server role to the free Microsoft Hyper-V Server R2. However, this process is a violation of license agreement, so StarWind urges the readers not to repeat it. As to the reason the process is at all possible, the answer is quite simple: SMB3 is a crucial part of the free Microsoft Hyper-V Server R2 and the latter won’t work if the protocol support is cut out. In any case, a fact that you can do something doesn’t mean you should. Violating Microsoft’s user license agreement is some serious business you don’t want to get involved in.

Microsoft Hyper-V Server 2012 R2 as the file server and Windows Server 2012 R2 as the client

Learn More

Posted by Anton Kolomyeytsev on January 22, 2016
Storage Spaces Direct: 4-node “Shared Nothing” Scale-Out File Server

This is a research dedicated to practical implementation of Microsoft Storage Spaces Direct. It is a part of a series of posts about S2D and features a detailed comprehensive instruction on building a fault-tolerant 4-node setup. Storage Spaces Direct is the next step of Storage Spaces, meaning it is an extension of the current SDS for Windows Server. Storage Spaces Direct utilizes SMB3 for all intra-node interaction, including the latest SMB Direct and SMB Multichannel. The testing of S2D is hot right now, so StarWind is doing its part as well. This setup is meant to withstand node failures and the post reveals how exactly it performs in this respect.

SOFS + Storage Spased Direct

Learn More

Posted by Anton Kolomyeytsev on November 20, 2015
Storage Spaces Direct: Overview

This is a short summary on Storage Spaces Direct – the first true software-defined storage from Microsoft, its history and peculiarities. Software-defined storage is a concept, which involves storing data without dedicated hardware. This is an introduction to a series of posts dedicated to theoretical and practical research of S2D. The post is not a full description of the Storage Spaces Direct technology, more info is available on Microsoft resources. Our focus here is the actual problem that S2D solved for virtualization world, being truly independent from underlying hardware, as SDS should. Its predecessor, Clustered Storage Spaces, had a very serious hardware lock-in with its requirements for SAS fabrics, SAS switches and SAS JBODS. Thus, Storage Spaces Direct is an interesting technology to research.


Learn More