June 23, 2016 at 15:39Dell Storage SC9000: Intelligent Storage for Efficient Data Center
The increase in storage performance requirements makes vendors look for new approaches to creating an optimal storage architecture and use flash memory along with traditional drives. The development of servers and storage systems is one of Dell's priorities, where the company invests heavily. Currently, it offers new backup and storage systems for various types of businesses. Let's take a closer look at the Dell SC9000 storage system .
The new Dell SC9000 disk array was released in October 2015 and definitely deserves a closer look. This flagship solution can already be seen in major data centers around the world. But the system is worth a look and medium-sized companies. SC9000 with multi-petabyte storage capacity and Storage Center 7.0 operating system offers new opportunities for IT departments.
The SC9000 hardware platform, the oldest in the SC series array of arrays, has the highest performance and scalability - up to 960 SAS SSD and / or HDD drives per array, and arrays can be combined into groups (federation).
The platform, which has already proven its reliability, is equipped with new Intel 8-core processors, the system memory capacity is quadrupled, and the SAS 12 Gb / s interface is used as a backend. The system supports SAN protocols (FC, iSCSI, FCoE), as well as NAS via the optional FS8600 device.
System performance has increased by 40% compared to previous SC products. Internal tests showed that the performance of the SC9000 array exceeds 360 thousand IOPS with a delay of less than a millisecond. In addition, throughput doubled, and this means the absence of bottlenecks with increasing load.
At the core of the Dell SC9000 is the new 13th-generation Dell PowerEdge server platform with four 3.2 GHz Intel Xeon 8-core processors and up to 512 GB of RAM. The system can be configured as a flash array or hybrid storage.
With redundant and hot-plug components, the SC9000 is a fail-safe and easy-to-deploy high availability solution. Expansion shelves SC400 / 420, SC200 / 220 or SC280 allow you to increase capacity. The system supports unified management with SC8000, SC4020, SCv2000 arrays.
Flash capacity is up to 46 TB per 1U (no data compression). This density was achieved thanks to the use of Samsung TLC 3D NAND three-level memory in storage systems. The SC9000 system provides more than 3 Pbytes of net capacity, and arrays can be grouped if workloads require large amounts of storage.
SC280 expansion shelves (and this is 96 3.5 ”HDDs) give a capacity of more than 100 TB per 1U. This can be used for media archives or in applications that work with large files.
Despite the impressive characteristics of the hardware platform, the SC9000 owes many of its qualities to software. The Storage Center virtualized software environment with automation features makes the system more economical and adequate for a wide range of tasks.
Comparison of software for arrays of the SC series.
With new features, Dell Storage Center 7.0 provides enhanced support for Dell Storage SC Series devices in private cloud environments and other mission-critical applications. Let's see what's new here.
Considering that the high cost of flash memory is a major barrier to adopting this technology, Dell has focused on improving drive efficiency in order to lower its cost. Many vendors claim optimization, but often a simplified approach is used with one level of storage, in which all types of flash memory behave the same.
Dell’s modern virtualization technologies enable you to use different types of flash in the same array to run different workloads and different application scenarios. Thanks to this, the company was able to release a series of high-speed and economical arrays both entirely on the basis of flash memory and in a hybrid version.
Dell auto-tiering takes into account different pricing / performance parameters for different types of SSDs. That is, to achieve high performance in IOPS you do not need to pay a prohibitive amount. That's why Dell flash arrays cost only 0.65 cents per gigabyte. Well, in hybrid storage systems this figure is even lower.
To increase the life of 3D NAND memory, a tearing mechanism is involved - moving data between storage levels.
The first level (Tier 1) is represented by SLC or eMLC flash memory that can withstand a large number of rewriting cycles, the second (Tier 2) is represented by 3D NAND memory, into which relatively rarely used data is written, and in a compressed form. In some models there is a third level (Tier 3) - HDD.
The system configuration (the ratio of the capacity of different types of flash memory and hard drives) can be selected taking into account the nature of the load: there is a corresponding utility for this. WI - for intensive recording, RI - for intensive reading
Dell's peering methods are among the most comprehensive in the industry. SC arrays allow you to create unified storage pools for SAN and / or NAS data from heterogeneous systems, to create volumes from different types of media (including including Tier 0 - server-side cache), and set up flexible RAID configurations. Due to the automatic placement of data at the appropriate level of storage in real time, the data “warming up” and “cooling” are accelerated. But the main thing is that management is simplified, and it is possible to squeeze everything out of the drives in terms of performance.
Intelligent data placement and storage optimization at all levels.
The patented technology for automatic multi-level data placement uses the main advantages of various types of drives to optimize data throughout the entire life cycle.
The same virtualization, which helps SC series arrays to demonstrate their price / performance advantages, allows new solutions to be introduced. A good example is the TLC 3D NAND. By launching Mainstream Read-Intensive drives last summer, Dell halved the cost of Tier 2 flash memory without sacrificing durability or performance. TLC memory is successfully used in its high-end arrays, while competitors are still trying to change the architecture of single-level systems of the lower class.
The simultaneous vertical and horizontal scalability makes SC arrays well extensible.
The flexible self-optimizing SC9000 system allows you to get more than 3 Pb of net capacity and combine arrays into groups, providing the necessary capacity for large workloads. The SC9000 provides fast modular expansion with support for Storage Area Networks (SAN) and Network Attached Storage (NAS). It can also be connected to other SC Series arrays in larger federated systems with a single control.
SCOS software includes the optional “storage hypervisor”, which makes it easy to move data across the “federation” of arrays. And even in the case of building one disk array, you can get the capacity and performance of a large distributed system with unified management.
The Dell Volume Advisor will provide proactive load balancing recommendations based on predefined rules. He monitors the federation of arrays and advises where it is best to place data.
SC Series arrays with SCOS 7.0 software support intelligent data compression on flash drives and HDDs. This results in significant capacity savings. The compression function is integrated with tearing, so compression does not reduce performance. At the same time, functions such as dynamic distribution of capacitance are active.
Data compression provides significant savings in storage capacity, which is especially important for flash memory.
Due to data compression, which reaches 93% in the case of Microsoft SQL databases, the solution becomes even more economical: the flash array is comparable in price to a disk array based on HDD SAS 15k.
Compression occurs when data blocks are moved to lower levels, thereby reducing the impact of data compression on system performance. When moving active blocks to the upper levels, they remain compressed.
A new SCOS 7.0 feature called Live Volume Auto-Failover prevents unplanned downtime. In the event of a failure, Storage Center redirects the load to the backup volumes synchronized with the primary volume on another array and starts the recovery procedure. Administrator intervention is not required.
Thanks to the reassignment of hosts, you can make changes to the storage without affecting the operation of users and applications.
Server-transparent movement of volumes between arrays in a federated cluster improves cache and capacity utilization under growing loads.
Another noteworthy SC9000 technology is the automatic replacement function of the failed Live Volume array, which allows you to work without downtime. It makes it possible to “stretch” volumes between platforms using the disk array itself. As a result, an inexpensive disaster-resistant solution is created that does not require restructuring the data storage architecture.
The Live Volume function with automatic failover ensures continuous operation of the system during power outages and automatically restores the high-availability environment when the array is turned on. No additional hardware or software is required.
The Live Volume feature on SC Series arrays is well-suited for cloud environments. It provides failover and transparently automatically switches to fully synchronized volumes on another SC Series array.
At the application level, snapshots are supported in the Oracle environment, cloud environments based on the Microsoft Azure platform, VMware Metro Stretch Clusters. The Azure Site Recovery feature backs up data to Microsoft's cloud infrastructure. It also provides integrated protection for Oracle, Microsoft, and VMware environments.
The Application Protection Manager software suite provides data integrity on the server in Oracle, VMware, and MS database environments. Optional FIPS-certified solid state drives and self-encrypting hard drives protect data from theft, loss or unauthorized access. And the built-in advanced features include comprehensive “subtle” methods of managing system resources.
In conclusion, compare the array with competitors:
The SC9000 array is the right solution for large data storage systems, demanding workloads, and distributed enterprise environments. The SC9000 array doesn’t just increase productivity: its intelligent virtualized architecture automatically optimizes financial costs and storage utilization, providing maximum savings.
The Dell SC9000 storage arrays are designed to help companies with large data centers more confidently and quickly achieve important goals while spending less time and money. Dell provides customers flexible, end-to-end solutions that can quickly adapt to changing needs, offering a solution with a record low cost per gigabyte of storage based on solid-state drives thanks to the new TLC 3D memory type.
Hardware platform
The new Dell SC9000 disk array was released in October 2015 and definitely deserves a closer look. This flagship solution can already be seen in major data centers around the world. But the system is worth a look and medium-sized companies. SC9000 with multi-petabyte storage capacity and Storage Center 7.0 operating system offers new opportunities for IT departments.
The SC9000 hardware platform, the oldest in the SC series array of arrays, has the highest performance and scalability - up to 960 SAS SSD and / or HDD drives per array, and arrays can be combined into groups (federation).
The platform, which has already proven its reliability, is equipped with new Intel 8-core processors, the system memory capacity is quadrupled, and the SAS 12 Gb / s interface is used as a backend. The system supports SAN protocols (FC, iSCSI, FCoE), as well as NAS via the optional FS8600 device.
System performance has increased by 40% compared to previous SC products. Internal tests showed that the performance of the SC9000 array exceeds 360 thousand IOPS with a delay of less than a millisecond. In addition, throughput doubled, and this means the absence of bottlenecks with increasing load.
At the core of the Dell SC9000 is the new 13th-generation Dell PowerEdge server platform with four 3.2 GHz Intel Xeon 8-core processors and up to 512 GB of RAM. The system can be configured as a flash array or hybrid storage.
With redundant and hot-plug components, the SC9000 is a fail-safe and easy-to-deploy high availability solution. Expansion shelves SC400 / 420, SC200 / 220 or SC280 allow you to increase capacity. The system supports unified management with SC8000, SC4020, SCv2000 arrays.
Flash capacity is up to 46 TB per 1U (no data compression). This density was achieved thanks to the use of Samsung TLC 3D NAND three-level memory in storage systems. The SC9000 system provides more than 3 Pbytes of net capacity, and arrays can be grouped if workloads require large amounts of storage.
SC280 expansion shelves (and this is 96 3.5 ”HDDs) give a capacity of more than 100 TB per 1U. This can be used for media archives or in applications that work with large files.
Software
Despite the impressive characteristics of the hardware platform, the SC9000 owes many of its qualities to software. The Storage Center virtualized software environment with automation features makes the system more economical and adequate for a wide range of tasks.
Comparison of software for arrays of the SC series.
With new features, Dell Storage Center 7.0 provides enhanced support for Dell Storage SC Series devices in private cloud environments and other mission-critical applications. Let's see what's new here.
Improved optimization for working with flash memory
Considering that the high cost of flash memory is a major barrier to adopting this technology, Dell has focused on improving drive efficiency in order to lower its cost. Many vendors claim optimization, but often a simplified approach is used with one level of storage, in which all types of flash memory behave the same.
Dell’s modern virtualization technologies enable you to use different types of flash in the same array to run different workloads and different application scenarios. Thanks to this, the company was able to release a series of high-speed and economical arrays both entirely on the basis of flash memory and in a hybrid version.
Dell auto-tiering takes into account different pricing / performance parameters for different types of SSDs. That is, to achieve high performance in IOPS you do not need to pay a prohibitive amount. That's why Dell flash arrays cost only 0.65 cents per gigabyte. Well, in hybrid storage systems this figure is even lower.
Deep tearing not limited to flash memory
To increase the life of 3D NAND memory, a tearing mechanism is involved - moving data between storage levels.
The first level (Tier 1) is represented by SLC or eMLC flash memory that can withstand a large number of rewriting cycles, the second (Tier 2) is represented by 3D NAND memory, into which relatively rarely used data is written, and in a compressed form. In some models there is a third level (Tier 3) - HDD.
The system configuration (the ratio of the capacity of different types of flash memory and hard drives) can be selected taking into account the nature of the load: there is a corresponding utility for this. WI - for intensive recording, RI - for intensive reading
Dell's peering methods are among the most comprehensive in the industry. SC arrays allow you to create unified storage pools for SAN and / or NAS data from heterogeneous systems, to create volumes from different types of media (including including Tier 0 - server-side cache), and set up flexible RAID configurations. Due to the automatic placement of data at the appropriate level of storage in real time, the data “warming up” and “cooling” are accelerated. But the main thing is that management is simplified, and it is possible to squeeze everything out of the drives in terms of performance.
Intelligent data placement and storage optimization at all levels.
The patented technology for automatic multi-level data placement uses the main advantages of various types of drives to optimize data throughout the entire life cycle.
Rapid introduction of new technologies
The same virtualization, which helps SC series arrays to demonstrate their price / performance advantages, allows new solutions to be introduced. A good example is the TLC 3D NAND. By launching Mainstream Read-Intensive drives last summer, Dell halved the cost of Tier 2 flash memory without sacrificing durability or performance. TLC memory is successfully used in its high-end arrays, while competitors are still trying to change the architecture of single-level systems of the lower class.
Fast vertical and horizontal scaling to meet storage needs and performance.
The simultaneous vertical and horizontal scalability makes SC arrays well extensible.
The flexible self-optimizing SC9000 system allows you to get more than 3 Pb of net capacity and combine arrays into groups, providing the necessary capacity for large workloads. The SC9000 provides fast modular expansion with support for Storage Area Networks (SAN) and Network Attached Storage (NAS). It can also be connected to other SC Series arrays in larger federated systems with a single control.
SCOS software includes the optional “storage hypervisor”, which makes it easy to move data across the “federation” of arrays. And even in the case of building one disk array, you can get the capacity and performance of a large distributed system with unified management.
The Dell Volume Advisor will provide proactive load balancing recommendations based on predefined rules. He monitors the federation of arrays and advises where it is best to place data.
Efficient use of capacity
SC Series arrays with SCOS 7.0 software support intelligent data compression on flash drives and HDDs. This results in significant capacity savings. The compression function is integrated with tearing, so compression does not reduce performance. At the same time, functions such as dynamic distribution of capacitance are active.
Data compression provides significant savings in storage capacity, which is especially important for flash memory.
Due to data compression, which reaches 93% in the case of Microsoft SQL databases, the solution becomes even more economical: the flash array is comparable in price to a disk array based on HDD SAS 15k.
Compression occurs when data blocks are moved to lower levels, thereby reducing the impact of data compression on system performance. When moving active blocks to the upper levels, they remain compressed.
Continuous availability
A new SCOS 7.0 feature called Live Volume Auto-Failover prevents unplanned downtime. In the event of a failure, Storage Center redirects the load to the backup volumes synchronized with the primary volume on another array and starts the recovery procedure. Administrator intervention is not required.
Thanks to the reassignment of hosts, you can make changes to the storage without affecting the operation of users and applications.
Server-transparent movement of volumes between arrays in a federated cluster improves cache and capacity utilization under growing loads.
Another noteworthy SC9000 technology is the automatic replacement function of the failed Live Volume array, which allows you to work without downtime. It makes it possible to “stretch” volumes between platforms using the disk array itself. As a result, an inexpensive disaster-resistant solution is created that does not require restructuring the data storage architecture.
The Live Volume function with automatic failover ensures continuous operation of the system during power outages and automatically restores the high-availability environment when the array is turned on. No additional hardware or software is required.
The Live Volume feature on SC Series arrays is well-suited for cloud environments. It provides failover and transparently automatically switches to fully synchronized volumes on another SC Series array.
Enterprise Application Integration
At the application level, snapshots are supported in the Oracle environment, cloud environments based on the Microsoft Azure platform, VMware Metro Stretch Clusters. The Azure Site Recovery feature backs up data to Microsoft's cloud infrastructure. It also provides integrated protection for Oracle, Microsoft, and VMware environments.
The Application Protection Manager software suite provides data integrity on the server in Oracle, VMware, and MS database environments. Optional FIPS-certified solid state drives and self-encrypting hard drives protect data from theft, loss or unauthorized access. And the built-in advanced features include comprehensive “subtle” methods of managing system resources.
In conclusion, compare the array with competitors:
The SC9000 array is the right solution for large data storage systems, demanding workloads, and distributed enterprise environments. The SC9000 array doesn’t just increase productivity: its intelligent virtualized architecture automatically optimizes financial costs and storage utilization, providing maximum savings.
The Dell SC9000 storage arrays are designed to help companies with large data centers more confidently and quickly achieve important goals while spending less time and money. Dell provides customers flexible, end-to-end solutions that can quickly adapt to changing needs, offering a solution with a record low cost per gigabyte of storage based on solid-state drives thanks to the new TLC 3D memory type.