Information Storage and Management: A Comprehensive Guide

Posted on Sep 10, 2024 in Technology

1. Active-Active vs. Active-Passive

AA-Symm

Hosts can perform I/Os through any controller.

AP-VNX

Hosts can perform I/Os only through the controller that owns the LUN (MODULE 4)

2. Advantages of IP (4)

• Uses existing network
• Easily scalable
• Reduced hardware cost
• Uses existing security options
• Uses existing long-distance recovery solutions

3. Application

Software program that provides logic for computing operations (MODULE 2)

4. Application Virtualization

Presenting an application to an end-user without installation or dependency on the user’s computing platform (MODULE 2)

5. Archive Architecture Components (3)

• Agent: Installed software on the application server scans data to be archived and leaves a stub reference file (small)
• Server: Software on the server defines archiving policies
• Storage Device: Stores fixed content

6. Backup Architecture (3)

• Backup Client: Gathers data to be backed up and sends it to the storage node
• Backup Server: Manages backup operations and keeps a catalog of backups
• Storage Node: Writes data to the storage device and manages it

7. Backup in Virtualized Environment (2)

• Traditional: Agent is located on the VM or on the hypervisor, taking precious processing power away from the host system
• Imaged-Based: Agent is on the hypervisor, takes a snapshot of the whole system, saves it as an image file, and saves it to the proxy server

8. Backup Methods (3)

• Hot: Application is running, and users may access data during backup
• Cold: Requires application shutdown
• Bare-Metal: Full system start from scratch

9. Backup Operation (7 Steps)

1. Backup server initiates backup
2. Retrieves information from the catalog
3. Backup server tells the node to tell storage to load the storage medium
4. Backup server tells clients to send data to the device
5. Clients send data
6. Node sends data to storage
7. Node sends metadata to the backup server
8. Backup server updates the catalog

10. Backup Targets (3)

• Tape: Portable, sequential, shoe-shining effect
• Disk: Random access, reliable
• Virtual Tape: Tricks existing data center software into thinking that they’re still using tape, not disk

11. Backup Topologies (4)

• Direct-Attached: Application server/client/node are all on the same device connected over LAN to the backup server and FC to the backup device
• LAN-Based Backup: Everything is separate and connected by a LAN
• SAN-Based Backup: FC used for everything except for LAN for metadata from the client to the backup server
• Mixed-Backup: Combination of LAN and SAN topologies

12. Benefits of OSD (4)

• Security and reliability (object ID)
• Platform independence
• Scalability
• Manageability (i.e., ATMOS)

13. Beta Methods to Make FCoE Lossless (Converged Enhanced Ethernet) (4)

• Priority-Based Flow Control (PFC): Uses pause functionality
• Enhanced Transmission Selection (ETS): Intelligently allocates bandwidth
• Congestion Notification (CN): Identifies congestion and notifies the host
• Data Center Bridging Exchange Protocol (DCBX): New advanced Ethernet protocol

14. Big Data

Data sets beyond the capacity of common software tools (i.e., Greenplum analytics software) (MODULE 1)

15. Block-Level Storage Virtualization

Using a virtual appliance, groups LUNs into massive pools of portable, non-disruptive virtual LUN volumes (i.e., EMC VPLEX)

16. Block vs. File Level Access

• Block: File system created on the host, requests for raw data sent over the network
• File Level: Request sent over the network to a separate file system, which gets data quickly but is more expensive (MODULE 2)

17. Business Continuity

Process that prepares for, responds to, and recovers from system outages (downtime)

18. Business Continuity Lifecycle (5 Steps)

1. Establishing objectives
2. Analyzing
3. Designing
4. Implementing
5. Training/Maintaining

19. Business Continuity Technology Solutions (3)

• Removing single points of failure with redundancy (failure of one component that terminates the availability of the entire system)
• Multipathing software (reroutes I/Os down different paths if one goes dead)
• Backup and replication

20. Business Impact Analysis

Identifies which business units are essential to the survival of the business and what a loss will cost a company

21. Cache

Large amount of volatile, semi-conductor memory. Grabs memory in nanoseconds (MODULE 4)

22. Cache Data Protection (2)

• Mirroring: Each write is sent to two memory locations
• Vaulting: Runs cache with battery until all memory is placed in vault drives (MODULE 4)

23. Cache Tiering

Uses DRAM for primary and FLASH drives for secondary cache memory

24. CAS Features (10)

• Content authenticity (uses binary blob to create ID)
• Location independence
• Single-instance storage (won’t copy the same object twice)
• Retention enforcement (for policy reasons)
• Data protection
• Faster retrieval
• Load balancing (multiple nodes)
• Scalability
• Self-diagnosis
• Audit-trail (i.e., CENTERA)

25. Cloud Computing

Model for convenient, on-demand access to a shared pool of computing resources that requires minimal management effort

26. Cloud Deployment Models (4)

• Public: Open use by everybody
• Private:
• Community: Intranet
• Hybrid: Some public/some private

27. COFW Principle vs. COFA Principle

COFW

Uses a bitmap that tracks changed blocks on the production FS – reads come from the production FS

28. Common I/O Characteristics of an Application (3)

• Read vs. write-intensive
• Sequential vs. random
• I/O size (MODULE 2)

29. Components of an Object (4)

• Object ID (unique algorithm ensures data integrity forever)
• Data
• Metadata
• Attributes

30. Components of iSCSI (3)

• Initiator (like host-based iSCSI HBA)
• Target (array or gateway)
• Network (IP)

31. Components of LVM

• Physical volumes (individual disks)
• Volume groups (groups of physical disks)
• Logical volumes (disk partitions) (MODULE 2)

32. Components of NAS Head

• CPU/Memory
• NIC
• OS
• Protocols
• Ports

33. Components of SAN (5)

• Node ports
• Cables
• Connectors
• Interconnecting devices
• Management software

34. Components of Unified Storage (4)

• Storage controller
• NAS head
• OSD node
• Storage

35. Compute Hardware Components (3)

• CPU
• Memory
• I/O devices (MODULE 2)

36. Compute (Host)

Resource that runs applications (MODULE 2)

37. Compute Software Components (4)

• OS
• Device drivers
• File system
• Volume manager (MODULE 2)

38. Compute Virtualization

Running multiple OS’s on one host by masking hardware components (MODULE 2)

39. Concatenation

LVM groups several disks to appear as one large disk (MODULE 2)

40. Connectors (3)

• SC-Standard
• LC-Smaller, better
• ST-Fiber patch channels

41. Content Address Storage (CAS)

Archiving solution designed to store fixed content

42. Continuous Data Protection Components (3)

Occurs at the network layer. Three components: journal volume, CDP appliance, and write splitter (i.e., EMC Recoverpoint)

43. Core Elements of a Data Center

“All Data Centers Seem Nice”

• Application
• Database
• Compute
• Storage
• Network (MODULE 1)

44. DAS

Direct Attached Storage. Storage attached directly to the host (internal or external). Pros: Cheap. Cons: Bad scaling and no storage pooling (MODULE 2)

45. Data

Collection of raw facts (MODULE 1)

46. Data Archive

A place where fixed content is stored: online, nearline, or offline

47. Data Deduplication

Process of identifying and reducing redundant data

48. Data Migration Replication

Moves data between heterogeneous

49. Data Transfer Rate (Internal/External)

Time taken for data to transfer from platter -> R/W head -> buffer -> interface -> HBA (MODULE 2)

50. DBMS

Database Management System. Organizes structured data (i.e., ORACLE, MYSQL). Application requests data, the database supplies it and tells the OS to get it from storage (MODULE 2)

51. Dedicated vs. Global Cache

• Dedicated: VNX separate memory for read and separate for writes
• Global: SYMM I/O from any memory location, great, only one set of addresses (MODULE 4)

52. Deduplication Implementations (2)

• Source-Based: Deduplicated at the source, less traffic over the network but increased overhead on the client
• Target-Based: Dedups on the target, offloading processing on the client but more network bandwidth used

53. Deduplication Methods (2)

• File Level or Single Instance: After one file is stored, all identical copies refer to it
• Subfile Level: More advanced, detects redundancy within files

54. Defense-in-Depth

Basically says just look everywhere while protecting

55. Define Backup and Uses (3)

Additional copy of data with the sole purpose of recovering lost data

• Disaster recovery
• Operational recovery
• Archive

56. Define iSCSI

Encapsulates SCSI I/O into IP packets and transports them with TCP/IP

57. Define NAS

IP-based, high-speed, dedicated file sharing and storage device

58. Define RAID

It is a technique that combines multiple disk drives into a logical unit (RAID set) and provides protection, performance, or both (MODULE 3)

59. Define SAN

High-speed, dedicated network of servers and shared storage devices. 15 million devices per network

60. Describe a Hot Spare

A hot spare refers to a spare drive in a RAID array that temporarily replaces a failed disk drive by taking the identity of the failed disk drive (MODULE 3)

61. Describe Dependent Write I/O Principle

If the writes from a database don’t all occur and in the correct order, the replica won’t make a copy

62. Describe FC-AL Connectivity Option

Nodes most share the network through a hub that supports up to 126 nodes or without a hub in a ring of devices

63. Describe FCIP Topology

Uses two parallel FCIP gateways that translate local FC data into IP packets for long-distance travel and convert them back into FC data at the other end

64. Describe FC-SW

Creates a logical space (fabric) in which all nodes communicate directly with each other through switches

65. Describe Flushing Host Buffers

Means the replica must flush memory in host RAM before it mirrors the data

66. Desktop Virtualization

Allows the whole user desktop environment (OS, user settings, apps) to be managed centrally and available dynamically to all types of devices (MODULE 2)

67. Device Driver

Software that enables the OS to recognize a specific device (MODULE 2)

68. Difference Between Backup and Replication

Replication is an exact copy that is instantly mirrored; in case of disaster, it offers seamless business continuance. However, backups are advantageous because they offer point-in-time historical data retrieval before user/environmental corruption occurred (http://bit.ly/yAiibn)

69. Disaster Recovery vs. Disaster Restart

• Disaster Recovery: All about RESTORING systems to where they were before the disaster, think backups
• Disaster Restart: All about RESTARTING business operations with replication technology. Think a McDonald’s burning down, disaster recovery is restoring the building to how it used to be before the fire, and disaster restart is the grand reopening where they start selling burgers again

70. Disk Buffered Remote Replication

Host writes data to the source, where it is locally replicated, which in turn is remotely replicated to another site, which is again locally replicated

71. Disk Drive Components (5)

• Platter
• Spindle
• R/W head
• Actuator arm
• Drive controller board (MODULE 2)

72. EMP

Enterprise Management Platform. It is a suite of applications that provides a great way to manage and monitor components

73. Essential Cloud Characteristics (5)

• On-demand self-service (need more devices/storage? Take them!)
• Broad network access (can use on several devices)
• Resource pooling
• Rapid elasticity (quickly scaled resources)
• Measured service (MR. ROB)

74. Fabric Services (4)

• Fabric login
• Name server
• Fabric controller (RSCN – “Hey all, node x joined the fabric, welcome him”)
• Management server

75. Fabric-Wide Access Control: Access Control Lists

Uses policy control, which specifies which HBAs and storage nodes can be connected to a particular switch

76. Fabric-Wide Access Control: Fabric Binding

Prevents an unauthorized switch from joining an existing switch in the network

77. Fabric-Wide Access Control: RBAC

Role-based access control enables the security admin to assign specific roles to different users

78. Factors for Digital Data Growth (4)

• Increased processing capabilities
• Low cost of storage
• Affordable communication technology
• Increasing number of smartphones and applications (MODULE 1)

79. Factors for Disk Drive Performance (3)

• Seek time
• Rotational latency
• Data transfer rate (MODULE 2)

80. Factors to Consider to Judge Vulnerability (3)

• Attack Surface: Refers to potential entry points
• Attack Vectors: Steps taken to launch an attack
• Work Factor: Refers to the amount of time and effort to exploit an attack vector

81. FC Addressing and Format

Assigned to nodes at fabric login, dynamic. Domain ID, Area ID, Port ID. 8 bits each. 239x256x256

82. FC Exchange

Occurs in level 4. Composed of one or more sequences. The protocol occurs between two hosts in an exchange

83. FC Frame

Fundamental transfer unit at level 2 that typically transfers SCSI data

84. FC Interconnecting Options (3)

• Point-to-point
• FC-AL (arbitrated loop)
• FC-SW (switched fabric)

85. FCIP

Long-distance solution, often associated with disaster recovery. This protocol allows virtual FC links across distributed FC data islands

86. FCoE

FCoE is a protocol that translates FC data over Ethernet, reducing cost and management

87. FCoE Components (3)

• Converged Network Adapter (CNA): Combination of FC HBA and standard NIC on the same card
• Cable: Copper and fiber optic
• FCoE Switch: Has Ethernet and FC switch capabilities

88. FC Protocol Stack (5)

• 0 – Defines the physical interface
• 1 – Defines how data is encoded
• 2 – Defines the structure of frames, routing, flow control, etc.
• 4 – Defines the protocols

89. FC SAN Topologies (4)

• Full Mesh: Each switch is connected to another switch; only one hop needed
• Partial Mesh: Not all switches are connected; multiple hops required
• Core-Edge Single: Switches attach to the director, which attaches to storage
• Dual Core-Edge: Same as single but with interconnecting directors

90. FC Sequence

Contiguous set of frames sent from one port to another

91. File

Collection of records/data stored as one unit (MODULE 2)

92. File-Level Virtualization

Virtualized appliance allowing a simple, non-disruptive, file-mobility solution by allowing logical paths to files instead of physical paths

93. File System

Hierarchical storage of files (i.e., FAT32, NTFS for Windows and UFS for Unix) (MODULE 2)

94. Fixed Content

Data that doesn’t change, at the end of its lifecycle (i.e., bank checks, x-rays)

95. Four Pillars of Multitenancy (4)

• Secure separation
• Service assurance
• Availability
• Management

96. Front End

Provides an interface between the host and storage system (MODULE 4)

97. Hardware vs. Software RAID Implementation

Uses host-based software, sacrifices CPU cycles, while hardware uses a controller on the host or array (MODULE 3)

98. Host-Based Remote Replication Methods (2)

• LVM-Based Replication: LVM sends data to two different sites
• Log Shipping: Database-related – everything is sent to a log (buffer) and periodically sent to the target

99. Host-Based Replication Methods (2)

• LVM-Based Mirroring: LVM writes each host write to two different physical locations
• File system snapshot

100. Hypervisor

Sits between VMs and hardware (MODULE 2)

101. ILM

Information Lifecycle Management, a proactive strategy to lower costs. Information is less valuable tomorrow than it is today

102. Information

Knowledge derived from data (MODULE 1)

103. Information Availability

Ability of IT infrastructure to function according to business expectations, during a specified time of operation. Uptime / (uptime + downtime)

104. Information Availability Defined by (3)

• Accessibility: Accessible to the user when required
• Reliability: Reliable and correct
• Timeliness: Window when it should be available

105. Information-Centric Storage

Environment where storage devices are managed centrally and shared across multiple servers (MODULE 1)

106. Information Security Framework (4)

• Confidentiality: Only authorized users have access
• Integrity: Information remains unaltered
• Availability: Users have reliable and timely access to systems
• Accountability: Log of events that ensure every process has an owner

107. Intelligent Storage System

RAID array with highly optimized I/O processing capabilities (MODULE 4)

108. Interconnecting Devices (3)

• Hubs
• Switches
• Directors

109. I/O Controller Percentage

Basically says that you shouldn’t use a controller past 70% of capacity, or you’ll lose performance (response time) (MODULE 2)

110. IOPS Problem Steps (2)

1. Determine disk capacity needed: total capacity / capacity of a single disk
2. Peak IOPS application needs / IOPS of disk
3. IOPS of disk = seek + (0.5 / disk rpm / 60) + (data block size / data transfer rate) (MODULE 2)

111. IP Protocols (2)

• iSCSI
• FCIP

112. iSCSI Security Implementations (2)

• CHAP: Challenge Handshake Authentication Protocol uses a hash
• iSNS: Similar to FC zoning, devices must be configured in the same domain

113. iSCSI Discovery Methods (2)

• SendTargets Discovery: Manually give the initiator the target’s information
• Internet Storage Naming Service (iSNS): Uses host-based software that auto-registers all targets

114. iSCSI Host Connectivity Options (3)

• Standard NIC with iSCSI software
• TCP Offload Engine (TOE): Like a NIC but offloads TCP processing to the card
• iSCSI HBA: Offloads all processing to the HBA

115. iSCSI Name (2)

Unique ID used to identify initiators and targets

1. iqn: Uses the organization’s registered domain name
2. eui: Burned-in global unique ID

116. iSCSI Topologies (3)

• Native: All components run on an IP-based network
• Bridged: Uses an iSCSI gateway to translate IP data to FC data
• Combination of native and bridged using an iSCSI and FC port on the array

117. ISL

Interswitch Link. Connection between switches. E-PORT

118. Key Characteristics of a Data Center

“PC MI ASS!”

• Performance
• Capacity
• Manageability
• Integrity
• Availability
• Security
• Scalability (MODULE 1)

119. Key Data Center Management Activities

• Monitoring
• Reporting
• Provisioning (MODULE 1)

120. Key Storage Components for Monitoring (3)

• Servers
• Networks
• Storage arrays

121. Local Replication in a Virtual Environment (2)

• Mirroring at the virtual volume
• Replication of virtual machines (clone & snapshot)

122. Local Replication Technologies (3)

• Host-based
• Network-based
• Storage array-based

123. Logical Volume Manager

Sits between the file system and physical disks. It can divide disks (partition) or join disks to appear bigger to the file system (concatenation)

124. Login Types in Switched Fabric (3)

• FLOGI: Between N&F “I would love to enter the fabric, may I? Yes, you may”
• PLOGI: Between N&N “May I introduce myself to you, fellow node?”
• PRLI: Between N&N “Let’s speak SCSI”

125. LRU vs. MRU

• LRU: Least Recently Used – ditches data that hasn’t been accessed for a long time (common)
• MRU: Most Recently Used – discards the most recently used data (MODULE 4)

126. LUN Masking

Occurs on the storage array. Decides which LUNs will be seen by which hosts (MODULE 4)

127. Magnetic Tape

Low-cost, sequential data solution for long-term storage (archiving) (MODULE 2)

128. Memory Virtualization

Fools the application into thinking there is more memory available than there is by using disk space and a swap-file process. Inactive memory is moved to disk and, when needed, swapped back to RAM (memory) (MODULE 2)

129. MetaLUN (2)

Is a method to expand LUNs.

• Concatenated: Provides capacity but no additional performance
• Striped: Adds capacity and performance (MODULE 4)

130. MTBF

Mean Time Between Failure. Calculations: total MTBF / # of drives (MODULE 3)

131. MTBF

Average uptime or time between failures. Total uptime / # of failures. The goal is 99.999%/year

132. MTTR

Average time to repair an outage. Total downtime / # of failures

133. Multitenancy

Multiple tenants using the same set of storage resources

134. Multi vs. Single-Mode Fiber Cables

• Multi-Mode: Only covers up to 500m because of modal dispersion
• Single-Mode: Up to 10km

135. NAS Backup Implementations (4)

• Server-Based: Bad – data is sent over LAN, overloading the network
• Serverless: Eliminates the client, one less hop, still over LAN
• NDMP 2-Way: Only metadata is sent over LAN, all else is sent off the network, good!
• NDMP 3-Way: Common private LAN for backup data. All implementations use NAS heads

136. NAS Device Components

• NAS head
• Storage

137. NAS File Sharing Protocols (2)

• CIFS: Stateful (auto-restores if connectivity is lost)
• NFS: Unix-based

138. NAS Implementations (3)

• Unified: Single storage platform for block/file, easy management
• Gateway NAS: Uses external/independent storage with a separate Gateway NAS head (allows FC in the back-end)
• Scale-Out: Isilon, pools multiple nodes together into a cluster for an easily scalable NAS device that pools resources

139. Object-Based Data

Unstructured data that’s at the end of its lifecycle stored as objects in flat-addressed space based on content, not address. Accessed over IP using REST and SOAP (XML-based)

140. Operating System

Software between applications and hardware responsible for controlling the entire environment (MODULE 2)

141. OSD System Components (3)

• OSD nodes; servers on the device that stores and assigns the object ID (metadata service) and maps storage locations (storage service) using an OS
• Internal network
• Storage

142. Parameters Managed (5)

• Capacity
• Reporting
• Accessibility
• Performance
• Security

143. Parameters Monitored (4)

• Capacity
• Accessibility
• Performance
• Security

144. Partitioning

LVM divides one disk into several logical disks (MODULE 2)

145. PCIe Card

Caching done on the host that uses smart technology (i.e., data used by the application frequently is kept on the host) (MODULE 4)

146. Physical Components of Connectivity (3)

• Host adapter (HBA or NIC)
• Port
• Cable (MODULE 2)

147. Physical Disk Structure

Platter -> tracks -> sectors (512 bytes) (MODULE 2)

148. Point-in-Time vs. Continuous

Point-in-time is timestamped like a backup, whereas continuous is in-sync with production data at all times

149. Point-to-Point Connectivity

Direct connection between nodes, used in a DAS environment

150. Port Types (4)

• N-Port: End of fabric, connects to node devices
• E-Port: Between two switches
• F-Port: On the switch that connects to the N-Port
• G-Port: Generic port on the back of the switch

151. Primary Cloud Service Models (3)

• Software-as-a-Service: Consumers deploy software, OS, and applications on the provider’s infrastructure (i.e., EC2)
• Platform-as-a-Service: Consumer-created apps are housed on the provider’s OS and infrastructure
• Infrastructure-as-a-Service: Consumers use the provider’s applications on the cloud infrastructure

152. Profiling

In a server-type backup, the process of taking a snapshot of application server CONFIGURATIONS

153. Protocol

Defined format for device communication done through controllers (i.e., IDE, ATA, SCSI) (MODULE 2)

154. RAID Array

Enclosure that contains physical and logical (RAID sets) drives with a RAID controller (MODULE 3)

155. Read Hit vs. Read Miss

• Read Hit: Host issues a request, the processor finds it in the cache
• Read Miss: Host issues a request, the processor gets it from disks, copies it to the cache, and sends it to the host (MODULE 4)

156. Remote Replication

Process of creating replicas at remote sites