EMC Isilon OneFS 7.0 Puts Enterprises in Front of Big Data Protection, Security and Virtualization Requirements
Bringing storage systems initially designed to meet Big Data demands into enterprise data centers is proving to be a bigger challenge than either storage providers or enterprises anticipated. While enterprises certainly want a storage system with a cost-effective, easy-to-manage, scalable architecture, other features such as data protection, data security and virtualization integration also come into play. EMC Isilon's latest OneFS 7.0 operating system takes these specific needs into account freeing enterprises to tackle their Big Data concerns.
Big Data Storage Systems
Storage systems designed to meet Big Data's ease of management, low cost and scalability demands got their start in high performance computing (HPC) environments with their architectures reflecting the demands that these environments placed on storage systems.
These HPC environments initially deployed low cost storage systems that used file systems such as the open source Lustre File System to deliver the high levels of performance these applications needed while enabling them to create and scale a single logical storage system. To do so required the use of proprietary, client-side drivers due to a very modest focus was put on providing scalable performance to industry standard file protocols such as NFS and CIFS.
EMC Isilon represents the evolution of these systems. It uses the OneFS file system and NFS, CIFS (SMB) and HDFS Hadoop protocols to provide the performance and scaling features that both the enterprise and commercial HPC environments expect without the need for client-side drivers to do network file sharing. Further, it provides multiple tiers of storage capacity that are all part of a single logical storage pool that can scale up to 15.5 petabytes (PBs) of raw storage capacity and handle 1.6 million CIFS file operations per second.
Maybe most importantly, it does not increase administrative overhead. More storage and performance capacity may be non-disruptively added to the system with the Isilon system offering the option to automate the placement of data on the most appropriate tier of storage.
These scale-out storage system features got the attention of enterprise organizations looking for better ways to store and manage their Big Data stores. However enterprises also have data protection, data security and virtualization needs that must be met before moving ahead with these next-generation storage systems. It is these concerns that the Isilon OneFS 7.0 release addresses.
Data Availability for All Circumstances
Data availability really takes two forms in storage systems. One is keeping data in a highly available state to ensure that uninterrupted, continuous data access in the event a storage system hard disk drive (HDD) or some other component fails. The second is creating reliable copies of data which may be used for restore should data corruption, data loss or interruption in service occur.
High Availability
The Isilon FlexProtect feature in the OneFS 7.0 release addresses data availability front by taking on one of the mounting concerns of traditional storage arrays: the time it takes using traditional RAID technology to recover data on a failed HDD. Due to the amount of data that resides on today's 1, 2 and 3 TB HDDs (and going to 4TB and beyond), another HDD in a RAID group could fail or an unrecoverable read error could occur before the data on the first failed HDD is restored.
The OneFS 7.0 file system deals with this by striping data across all nodes in a cluster and provide complete data availability with up to 4 node failures or 4 drive failures in a single node (n+4 data protection.) Should an HDD or a node fail, data reconstruction times are accelerated as the information needed for the recovery is distributed and stored on each of the nodes in the cluster, with all nodes working in parallel and participating in the reconstruction process.
Protecting data this way also has two other side benefits. First, it serves to increase the systems performance since data may be accessed or written to all of these HDDs or nodes in parallel. Second, its hot spares are virtual so if an HDD or node fails, Isilon can reconstruct data on a virtual hot space that consists of available capacity across its entire system.
Disaster Recovery and High Availability
To back data up, Isilon's SnapshotIQ feature can take as frequently as needed an unlimited number of snapshots at the directory and even the file level. This flexibility gives organizations the option to set short snapshot intervals for specific files or directories that have rapidly changing data.
The Isilon FileRevert, SnapRevert and SyncIQ features works in hand in glove with its SnapshotIQ feature. The FileRevert and SnapRevert automates the restoration of entire files and snapshots respectively while its SyncIQ feature may be used to first replicate data to another Isilon cluster and then automates the failover/failback processes between the two clusters simplifying disaster recovery operations.
One Storage System, Multiple Security Options
Data security is like data availability in that it also must satisfy multiple organizational requirements. Organizational data security needs are typically driven by three separate forces: meeting internal and external compliance requirements, maintaining internal user roles and logically segregating data and storage. The OneFS 7.0 release continues to build upon and enhance previously existing features to meet these specific organizational requirements.
Isilon Responds to External Compliance and Internal Threats
Its SmartLock™ feature now includes customized default retention periods for specific regulations. Turning this feature on ensures that data is retained for an appropriate length of time so organizations satisfy established regulatory standards such as are included in SEC 17a-4. It also ensures that data is kept for exactly the right length of time by accounting for events such as leap years and even time zone changes when calculating data retention periods.
A means to set policies is only equaled in importance by the ability to ensure only the appropriate individuals can access the data as internal vulnerabilities now account for about 70 percent of all data breaches. To ensure only the right individuals access data at the right time with the appropriate privileges, Isilon has beefed up its Role Based Administration feature to provide two modes: Enterprise mode and Compliance mode.
Managing data in Enterprise mode will give authorized administrators the flexibility to delete data on the system. Compliance mode is the more stringent of the two as it disables login by root users. This protects against either accidental deletion or malicious modification of both data and metadata. More importantly, by using the Compliance mode organizations get the security features that they need to make SEC17a-4 compliance a reality.
One System, Multiple Storage Pools
Isilon wisely kept the roles of data and storage management separate. In this way, organizations can create and manage a single logical storage system that is managed by one person or group but which can host multiple secure isolated storage pools using Isilon's Authentication Zones feature. In this way each storage pool may have its own administrator to keep their data secure even as the organization as a whole consolidates its data on Isilon to optimize available storage capacity.
Virtualization with an Eye on Automation
Gartner forecasts that total virtual OS instances will reach 70.2% of total OS instances by 2012 and 82.4% by 2016. This growth in virtualization adoption almost dictates that any storage array an organization brings into their infrastructure needs to support server virtualization in general and VMware specifically.
The OneFS 7.0 release ensures that Isilon is a fit in today's virtualized environment. In addition to providing support for the VMware vStorage APIs for Array Integration (VAAI) and the VMware vStorage APIs for Storage Awareness (VASA), Isilon with this release gets in front of the virtualization curve by providing platform APIs for independent software vendors (ISV).
The addition of these platform-specific APIs is significant in that organizations are looking to do more than simply virtualize their infrastructure - they are looking to automate it as well. This automation becomes possible with the introduction of these APIs. Initially available as a REST-based HTTP, organizations who take advantage of these APIs will get new flexibility to automate, control and provision storage in an Isilon cluster.
EMC Isilon OneFS 7.0 Ready to Meet Big Data and Business Demands
Enterprises are ready to simplify the storage management and overhead associated with their growing Big Data stores and workloads. However they only want to do so once they are confident the storage system used to support their Big Data stores can provide the storage architecture that Big Data stores demand and the security, protection and virtualization features that businesses need.
The Isilon OneFS 7.0 release brings these two sometimes competing priorities together in a single storage system. Isilon has always had a cost-effective, scalable storage system. But now with its OneFS 7.0 release, Isilon rounds out its data protection, data security and virtualization options so that organizations have the full suite of features that they need to confidently deploy it in their environment.
Big Data Storage Systems
Storage systems designed to meet Big Data's ease of management, low cost and scalability demands got their start in high performance computing (HPC) environments with their architectures reflecting the demands that these environments placed on storage systems.
These HPC environments initially deployed low cost storage systems that used file systems such as the open source Lustre File System to deliver the high levels of performance these applications needed while enabling them to create and scale a single logical storage system. To do so required the use of proprietary, client-side drivers due to a very modest focus was put on providing scalable performance to industry standard file protocols such as NFS and CIFS.
EMC Isilon represents the evolution of these systems. It uses the OneFS file system and NFS, CIFS (SMB) and HDFS Hadoop protocols to provide the performance and scaling features that both the enterprise and commercial HPC environments expect without the need for client-side drivers to do network file sharing. Further, it provides multiple tiers of storage capacity that are all part of a single logical storage pool that can scale up to 15.5 petabytes (PBs) of raw storage capacity and handle 1.6 million CIFS file operations per second.
Maybe most importantly, it does not increase administrative overhead. More storage and performance capacity may be non-disruptively added to the system with the Isilon system offering the option to automate the placement of data on the most appropriate tier of storage.
These scale-out storage system features got the attention of enterprise organizations looking for better ways to store and manage their Big Data stores. However enterprises also have data protection, data security and virtualization needs that must be met before moving ahead with these next-generation storage systems. It is these concerns that the Isilon OneFS 7.0 release addresses.
Data Availability for All Circumstances
Data availability really takes two forms in storage systems. One is keeping data in a highly available state to ensure that uninterrupted, continuous data access in the event a storage system hard disk drive (HDD) or some other component fails. The second is creating reliable copies of data which may be used for restore should data corruption, data loss or interruption in service occur.
High Availability
The Isilon FlexProtect feature in the OneFS 7.0 release addresses data availability front by taking on one of the mounting concerns of traditional storage arrays: the time it takes using traditional RAID technology to recover data on a failed HDD. Due to the amount of data that resides on today's 1, 2 and 3 TB HDDs (and going to 4TB and beyond), another HDD in a RAID group could fail or an unrecoverable read error could occur before the data on the first failed HDD is restored.
The OneFS 7.0 file system deals with this by striping data across all nodes in a cluster and provide complete data availability with up to 4 node failures or 4 drive failures in a single node (n+4 data protection.) Should an HDD or a node fail, data reconstruction times are accelerated as the information needed for the recovery is distributed and stored on each of the nodes in the cluster, with all nodes working in parallel and participating in the reconstruction process.
Protecting data this way also has two other side benefits. First, it serves to increase the systems performance since data may be accessed or written to all of these HDDs or nodes in parallel. Second, its hot spares are virtual so if an HDD or node fails, Isilon can reconstruct data on a virtual hot space that consists of available capacity across its entire system.
Disaster Recovery and High Availability
To back data up, Isilon's SnapshotIQ feature can take as frequently as needed an unlimited number of snapshots at the directory and even the file level. This flexibility gives organizations the option to set short snapshot intervals for specific files or directories that have rapidly changing data.
The Isilon FileRevert, SnapRevert and SyncIQ features works in hand in glove with its SnapshotIQ feature. The FileRevert and SnapRevert automates the restoration of entire files and snapshots respectively while its SyncIQ feature may be used to first replicate data to another Isilon cluster and then automates the failover/failback processes between the two clusters simplifying disaster recovery operations.
One Storage System, Multiple Security Options
Data security is like data availability in that it also must satisfy multiple organizational requirements. Organizational data security needs are typically driven by three separate forces: meeting internal and external compliance requirements, maintaining internal user roles and logically segregating data and storage. The OneFS 7.0 release continues to build upon and enhance previously existing features to meet these specific organizational requirements.
Isilon Responds to External Compliance and Internal Threats
Its SmartLock™ feature now includes customized default retention periods for specific regulations. Turning this feature on ensures that data is retained for an appropriate length of time so organizations satisfy established regulatory standards such as are included in SEC 17a-4. It also ensures that data is kept for exactly the right length of time by accounting for events such as leap years and even time zone changes when calculating data retention periods.
A means to set policies is only equaled in importance by the ability to ensure only the appropriate individuals can access the data as internal vulnerabilities now account for about 70 percent of all data breaches. To ensure only the right individuals access data at the right time with the appropriate privileges, Isilon has beefed up its Role Based Administration feature to provide two modes: Enterprise mode and Compliance mode.
Managing data in Enterprise mode will give authorized administrators the flexibility to delete data on the system. Compliance mode is the more stringent of the two as it disables login by root users. This protects against either accidental deletion or malicious modification of both data and metadata. More importantly, by using the Compliance mode organizations get the security features that they need to make SEC17a-4 compliance a reality.
One System, Multiple Storage Pools
Isilon wisely kept the roles of data and storage management separate. In this way, organizations can create and manage a single logical storage system that is managed by one person or group but which can host multiple secure isolated storage pools using Isilon's Authentication Zones feature. In this way each storage pool may have its own administrator to keep their data secure even as the organization as a whole consolidates its data on Isilon to optimize available storage capacity.
Virtualization with an Eye on Automation
Gartner forecasts that total virtual OS instances will reach 70.2% of total OS instances by 2012 and 82.4% by 2016. This growth in virtualization adoption almost dictates that any storage array an organization brings into their infrastructure needs to support server virtualization in general and VMware specifically.
The OneFS 7.0 release ensures that Isilon is a fit in today's virtualized environment. In addition to providing support for the VMware vStorage APIs for Array Integration (VAAI) and the VMware vStorage APIs for Storage Awareness (VASA), Isilon with this release gets in front of the virtualization curve by providing platform APIs for independent software vendors (ISV).
The addition of these platform-specific APIs is significant in that organizations are looking to do more than simply virtualize their infrastructure - they are looking to automate it as well. This automation becomes possible with the introduction of these APIs. Initially available as a REST-based HTTP, organizations who take advantage of these APIs will get new flexibility to automate, control and provision storage in an Isilon cluster.
EMC Isilon OneFS 7.0 Ready to Meet Big Data and Business Demands
Enterprises are ready to simplify the storage management and overhead associated with their growing Big Data stores and workloads. However they only want to do so once they are confident the storage system used to support their Big Data stores can provide the storage architecture that Big Data stores demand and the security, protection and virtualization features that businesses need.
The Isilon OneFS 7.0 release brings these two sometimes competing priorities together in a single storage system. Isilon has always had a cost-effective, scalable storage system. But now with its OneFS 7.0 release, Isilon rounds out its data protection, data security and virtualization options so that organizations have the full suite of features that they need to confidently deploy it in their environment.
Leave a comment