RAID Storage Architecture (Cont’d) PDF

Certified Cybersecurity Technician Exam 212-82 Data Security RAID Storage @ P Architecture (Cont’d).....

Certified Cybersecurity Technician Exam 212-82 Data Security RAID Storage @ P Architecture (Cont’d).. ’ i0.2t CC:)‘/\/ v Cache is used to write v" nvSRAM is the fastest v" Provides a v' The hardware presents the data in ! in for the RAID the RAID to the host A RAID system uses a the industry with 20 system's primary system as a single and cache to speed up I/O ns read and write cache performance on the access time storage system v | All Rights Reserved. Reproduction is Strictly Prohibited. RAID Storage Architecture The RAID architecture depends on two principles: redundancy and parallelism, providing a wide range of storage options with better performance and freedom from disk failures. The Internet’s increasing footprint has led to an increase in the use of RAID systems because of their high data storage capabilities and management systems. Many available RAID implementations depend on the following factors: parallelism, duplication, and redundancy. In a RAID architecture, a switch receives the data from servers connected to the network. The switch then sends the data to the processor at a later stage. The processor transfers the received data to a RAID controller. The RAID controller may be implemented either as a hardware using a RAID-on-Chip (ROC) or in a software. The ROC can contain the 1/0 interfaces, a processor, a host interface, and a memory controller. The ROC is installed directly in a motherboard using an expansion card, or in an external drive enclosure. Module 15 Page 1861 Certified Cybersecurity Technician Copyright © by EG-Council All Rights Reserved. Reproduction is Strictly Prohibited. Certified Cybersecurity Technician Exam 212-82 Data Security Configuration Network Connection to Host Data y wnmedl [ B OO D B I T Y | SDRAM l Q RAID RAID Journal, :’, Controller Transaction, -SRI “‘H SAN/NAS/Host - and Error Log ~ DIMM Interface T File é A A Multiport G dh sl. ST |IR Fo— » Memory Packup Coy é Y Controller HES! Processor RAID Control Processor SAS/SATA Interfaces — 13,| NAND | A: — FLASH Y Battery Backup or Ultracapacitor Unit..... | Primary RAID Memory Cache Figure 15.86: RAID storage architecture The RAID storage architecture outlines how the RAID server functions. The processor controls the entire functioning of the drive arrays and interfaces. It provides flexible and high- performance functions. The architecture in the figure above shows a RAID system can depend on hard disk drives (HDDs) as well as SSDs. The processor requires DRAM and NAND flash memory. The NAND flash memory provides a nonvolatile storage to the primary RAID memory cache. A battery backup or an ultra-capacitor unit in the primary RAID memory cache is helpful when the RAID control processor suffers from a power failure. In this scenario, the battery backup independently copies the DRAM contents to the NAND flash memory. A battery backup is an inexpensive alternative during a power loss. The architecture shows the requirement of a nonvolatile memory in the RAID controller firmware, RAID journal, and transaction and error log files. Major Components of a RAID Architecture = RAID controller: This is either hardware- or software-based and contains the HDDs or solid state drives as a single logical unit. A RAID controller has the permission to access multiple copies of files present on multiple disks, thereby preventing damage and increasing the system performance. In a hardware RAID, a physical controller manages the RAID array with a controller in the form of a PCl card that supports SATA or SCSI. A software RAID works similarly to a hardware RAID, except that their performance is lower than the former. = Primary RAID memory cache: The RAID controller has a direct access to the cache memory, enabling faster read and write access to the storage system. The cache is used Module 15 Page 1862 Certified Cybersecurity Technician Copyright © by EG-Gouncil All Rights Reserved. Reproduction is Strictly Prohibited. Certified Cybersecurity Technician Exam 212-82 Data Security to store the changing data. The cache memory is bigger in size and uses high-speed SDRAMs. A normal cache memory has a write cache and a separate read cache. The read cache decreases the latency of the read process. There are two types of write cache memories: o Write-through mode: This bypasses the cache memory and writes the data directly to the disk after the host sends it. The host sends the next data item after receiving a confirmation that the writing process has been completed. o Write-back mode: Data sent from the host is written to the cache memory. The host may perform other actions while the RAID controller transfers data from the cache to the disk drive. The RAID controller acknowledges the write process to the host soon after writing the data to the cache. Issues may arise if a RAID controller sends an acknowledgment before the data has been completely written on the disk. = |DE, SATA, or SCSI interfaces: IDE, SATA, or SCSI are device cables that transmit read/write signals to and from the drive. These are mostly used for internally connecting the drives. Moreover, servers are connected using these interfaces. o IDE: Integrated drive electronics (IDE) allows the connection of two devices per channel. It is normally used for internal devices as the cables are large and flat. o SATA: Serial ATA deals with hot plugging and serial connectivity. The hot plugging technique may be used to replace computer components without shutting down the system. SATA enables only one connection per connector and is not flexible for industrial purposes. o SCSI: Small computer system interface (SCSI) allows multiple devices to be connected to a single port at the same time. SCSI uses a parallel cable for attaching internal and external devices. = nvSRAM: Nonvolatile SRAM, or nvSRAM, has a faster read and write process (20 ns read and write access time) because of the presence of a standard asynchronous SRAM interface. nvSRAM ensures adequate data storage capabilities without the need for a battery during a shut down. nvSRAM is best used in applications that require high speed and nonvolatile storage at a low cost, such as in the medical industry. nvSRAM backups the data even in the event of a power failure. = Multiport memory controller: An MPMC provides access to memory for up to eight ports. A memory controller can be present as a separate chip or as an integrated memory. It provides access to a memory bank, and helps achieve a high efficiency with random address accesses. = NAND flash memory: Flash memory is a storage medium designed from electrically erasable programmable read-only memory (EEPROM). NAND and NOR are two types of flash memories. It provides a nonvolatile storage for the RAID system's primary cache. Its primary aim is to reduce cost and increase capacity. It does not require power to retain the data. It can improve its read-write cycles with reduced voltage demands. Module 15 Page 1863 Certified Cybersecurity Technician Copyright © by EG-Council All Rights Reserved. Reproduction is Strictly Prohibited. Certified Cybersecurity Technician Exam 212-82 Data Security = SDRAM: Synchronous dynamic random access memory or synchronous DRAM (sDRAM) is @ memory that is synchronized with the processor’s clock speed. This increases the number of instructions the processor can process. SDRAM speed is measured in Mega Hertz (MHz). It is divided into several sections called banks that allow the device to operate on several memory access commands simultaneously. = Disk: The hardware presents the RAID to the host system as a single and large disk. Module 15 Page 1864 Certified Cybersecurity Technician Copyright © by EG-Gouncil All Rights Reserved. Reproduction is Strictly Prohibited. Certified Cybersecurity Technician Exam 212-82 Data Security RAID Level 0: Disk Striping c RAID Level 0 splits data into blocks that are written evenly across multiple hard drives " It improves 1/0 performance by spreading the 1/0 load across several channels and disk drives Data recovery is not possible if a drive fails fi It requires a minimum of two drives It does not provide data redundancy RAID Level 0: Disk Striping Depending on the requirement of your organization, you can choose any RAID level. RAID levels are based on performance, fault tolerance, or both. RAID 0 deals with data performance. In this level, data is broken into sections and written across multiple drives. The storage capacity of RAID 0 is equal to the sum of the disks’ capacities in the set. RAID 0 does not provide fault tolerance. It requires a minimum of two drives. It does not provide data redundancy. A failure of one disk can lead to the failure of all disks in a level 0 volume. The probability of recovering data from a RAID level 0 is minimal. The data distribution in a RAID level 0 is equal among all the disk sets, resulting in high performance. With concurrent high performance, the throughput of the read and write operations on multiple disks is equal to the throughput of the array of disks. Increased throughput is an advantage of RAID 0, considering that data recovery is not available. Software and hardware RAID controllers support RAID 0, helping boost server performance. Example: Assume that the IT infrastructure has a hard disk with high performance. The data in the hard disk is transferred at a remarkably high speed. All the large and critical files are stored in this disk. However, if this disk fails, the entire contents of the files will be affected, leading to the unavailability of the data. It is advisable not to store any critical data in a RAID level 0. Advantages of RAID Level 0 = Read and write performance: RAID level 0 has a good read and write performance. The performance is even better when the controller supports independent reads and writes to different disks in the array. = Cost: RAID level 0 is more cost effective than the other RAID levels. Module 15 Page 1865 Certified Cybersecurity Technician Copyright © by EG-Gouncil All Rights Reserved. Reproduction is Strictly Prohibited. Certified Cybersecurity Technician Exam 212-82 Data Security * Implementation: Is easy to implement as the data is divided in a sequential set of blocks. There is no storage loss as the maximum capacity is utilized. Disadvantages of RAID Level 0 * No redundancy: With no data redundancy, data loss is greater. = Noncritical data: Data that is not critical to the organization can be stored on RAID level 0. This level does not use mirroring. Recovery is not possible if critical data is lost on RAID level 0. = Unreliable: If one disk fails, the entire network will be affected. RAID 0 T Disk 1 Figure 15.87: RAID level 0 Module 15 Page 1866 Certified Cybersecurity Technician Copyright © by EG-Council All Rights Reserved. Reproduction is Strictly Prohibited.