Classic Data Centre (CDC)

Meaning: A CDC is a facility containing physical IT resources including compute, network,

and storage
Core Elements of CDC:

- Application
- DBMS
- Compute
- Storage
- Networking

Key Requirements of a Data Centre:

A data center solution that considers and designs for the five key elements; performance,
time, space, experience and sustainability, will be reliable, flexible, scalable and efficient
in many ways beyond just cooling and power

Applications:
• Commonly deployed applications in a CDC

• Business applications

• E-mail, Enterprise Resource Planning (ERP), Decision Support System

(DSS), Data Warehouse (DW)

• Management applications

• Resource management, performance tuning

• Data protection applications

• Backup, replication

• Security applications

• Authentication, antivirus
• Key I/O characteristics of an application

• Read intensive vs. write intensive

• Sequential vs. random

DataBase Management Systems (DBMS):
• Database is a structured way to store data in logically organized tables that are
interrelated

• Helps to optimize the storage and retrieval of data

• DBMS is a collection of computer programs that control the creation, maintenance,
and use of databases

• Processes an application’s request for data

• Instructs the OS to retrieve the appropriate data from storage

• Popular DBMS examples are MySQL, Oracle RDBMS, SQL Server, etc.

Compute:
A resource that runs applications with the help of underlying computing components
• Compute consists of physical components (hardware devices) and logical components
(software and protocols)
• Physical components of compute are CPU, Memory, and Input/Output (I/O) devices
• I/O devices facilitate the following types of communication:

• User to compute: Handled by basic I/O devices such as keyboard, mouse, etc.

• Compute to compute/storage: Enabled using host controller or host adapter

Examples of Compute Systems:
• Examples of compute systems:

• Laptops/Desktops

• Blade servers

• Complex cluster of servers

• Mainframes
• Bladed server technology is commonly used to deploy compute systems in a CDC

• Consolidates power- and system-level function into a single, integrated chassis

• Enables the addition of server modules as hot-pluggable components

• Provides increased server performance and availability without increase in

size, cost, or complexity
Logical Components of Compute:
Storage:
It is a resource that stores data persistently for subsequent use.
• Data created by individuals/businesses must be stored for further processing
• The type of storage device used is based on the type of data and the rate at which it is
created and used
• A storage device may use magnetic, optical, or solid state media

• Examples: Disk drive (magnetic), CD (optical), Flash drive (solid state)

Storage Device Options:

Tape Drive
• Low cost solution for long term data storage
• Sequential data access, physical wear and tear, and storage/retrieval overheads
Optical Disk
• Write Once and Read Many (WORM): CD, DVD
• Limited in capacity and speed
Disk Drive
• Random read/write access
• Uses mechanical parts for data access
Solid State Drive
• Provides ultra high performance required by mission-critical applications
Intelligent Storage Systems:
• Is a RAID array highly optimized for I/O processing
• Have large amounts of cache for improving I/O performance
 • Have operating environments that provide:

• Intelligence for managing cache

• Optimal management, allocation, and utilization of storage resources

RAID: A technology which utilizes multiple disk drives as a set to provide protection,
capacity, and/or performance benefits.
Benefit: Overcomes limitations of disk drives
Improves storage system performance
- By serving I/Os from multiple disks simultaneously
RAID techniques are:
Striping, mirroring, and parity

Components of Intelligent Storage Systems:

Presenting Storage to Compute Systems:

Storage Networking Technologies:

 • Compute to compute communication
• Compute to storage communication
Direct Attached Storage (DAS)
Fibre Channel SAN (FC-SAN)

Compute to compute communication

• Typically uses Ethernet or TCP/IP protocol

• LAN, MAN, and WAN

• Communication is enabled using various components:

• Network Interface Card (NIC)

• Has unique MAC address

• Switches and routers

• Switch provides scalability and interconnection between multiple

compute systems

• Routers allow different networks to communicate with each other

• Cables
Twisted pair, co-axial cable, optical fiber

Compute to storage communication

• Communication is enabled using various hardware components (HBA, CNA,
NIC, switch, router, gateway ,and cables) and protocols
• Communication between compute and storage can be done using channel or network
technologies
 Communication Protocols:
• Peripheral Component Interconnect (PCI)

• Provides interconnection between CPU and attached devices

• Latest PCI Express bus provides throughput of 133 MB/sec

• Integrated Device Electronics/Advanced Technology Attachment (IDE/ATA)

• Popular protocol to connect to disk drives

• Supports 16-bit parallel transmission

• Serial version is called Serial ATA (SATA)

• Both versions offer good performance at a relatively low cost

• Small Computer System Interface (SCSI)

• Preferred storage connectivity option for high-end environments

• Improved performance, scalability, and high cost when compared to ATA

• Serial version is called Serial Attached SCSI (SAS)

• Transmission Control Protocol/Internet Protocol (TCP/IP)

• Traditionally used for compute to compute communication

• Now used for compute to storage communication also

• iSCSI (SCSI over IP) and FCoE (Fibre Channel over Ethernet) are
examples
Data Access by Compute:

Direct Attached Storage (DAS): An internal or external storage device, which

connects directly to a compute system.
• DAS is classified as internal or external based on the location of the storage device
with respect to the compute system
• Benefits:

• Simple to deploy and ideal for local data provisioning

• Low capital expense and less complexity

• Challenges:

• Limited scalability

• Limited ability to share resources

• Islands of over and under-utilized storage resources

Emergence of Storage Networking Technologies:
• Just-in-time information for business users
• Flexible and resilient storage architecture
• DAS is inefficient to fulfill these requirements
• Storage networking technologies emerged as a solution
• Fibre Channel SAN (FC SAN)
• Network Attached Storage (NAS)
• Internet Protocol SAN (IP SAN)
• Fibre Channel over Ethernet (FCoE)
• Object Based storage
• Unified storage

Object Based and Unified Storage Technologies

Object Based Storage:
• Object Based Storage combines data with rich metadata to create an “object”
• Object Based Storage stores data in a flat address space

• There are no hierarchies or nested directories

• Each object is identified by an unique ID (Object ID)

• Generated by a hashing function

• Object Based Storage uses HTTP communication as its standard interface

• SOAP and REST are the protocols commonly used in object based
communication in Cloud
• Simple Object Access Protocol (SOAP ) is used for communication between peers in
a distributed environment

• Uses Extensible Markup Language ( XML) framework

• Representational State Transfer (REST) is used to retrieve information from a
Website by reading Web pages
Why Object Base Storage?
Increasing amount of unstructured data

• SAN is highly scalable and supports data access at a block level

• Not a good option for data sharing

• NAS is a good option for applications which need to share data

• Limited scalability due to hierarchical structure

• Object Based approach potentially eliminates SAN and NAS limitations

• Highly scalable with data sharing capabilities

Benefits of Object Storage:

• Automates and simplifies storage management
• Ensures data integrity
• Ensures compliance and auditability
• Enables easy data migration
• Enables self healing
• Facilitates intelligent replication
• Allows flexible scalability

• Unified Storage: Provides Consolidated Management interface for NAS, SAN,

iSCSI, FCoE, and object based technologies.

Benefits of Unifies Storage:

• Provides consolidated multi-protocol storage

• File: NFS, CIFS

• Block: iSCSI, FC, FCoE

• Object: REST, SOAP

• Simplifies administration

• Integrated management interface

• Reduces cost of storage assets, along with power, cooling, and space
Provides a highly scalable architecture
=============== =======================

Replication:

• Process of creating an exact copy of data

• Drivers for replication

• Alternate source for backup

 • Fast recovery

• Decision support

• Testing platform

• Restart from replica

• Classification of Replication:

• Local replication

• Remote replication

Types and Characteristics:

• Types of Replica: Choice of replica ties back to RPO

• Point-in-Time (PIT)

• Non-zero RPO

• Continuous

• Near-zero RPO
• Characteristics of a good replica:

• Recoverability

• Replica should be able to restore data on the source device

• Restartability

• Restart business operation from replica

• ConsistencyTypes and Characteristics

• Consistent replica ensures that the data buffered in the compute system
is properly captured on the disk when the replica is created
Types:
Local Replication:
• Process of replicating data within the same array or the same data center
• Compute based replication

• Replication is performed by using CPU resources of the compute system

 • Types: LVM based mirroring and File system Snapshot
• Storage array based replication

• Replication is performed by using CPU resources of the storage array

• Types of Storage array based replication techniques:

• Full volume mirroring

• Pointer based full volume replication

• Pointer based virtual replication

Compute Based Replication:
• Logical Volume Manager based mirroring

• Each logical partition in a logical volume is mapped to two physical partitions

on two different physical volumes

• Write to a logical partition is written to the two physical partitions

• File System Snapshot

• Pointer-based local replication uses Copy on the First Write (CoFW) principle

• Uses bit map and block map

• Requires a fraction of the space used by the production FS

Pointer Based Full Volume Replication:
Full Volume Mirroring
• Target is a full physical copy of the source device
• Target is attached to the source and data from the source is copied to the target
• Target is unavailable while it is attached
• Target device is as large as the source device

• Provides a full copy of the source data on the target

• Target device is made accessible for business operations as soon as the replication
session has started
 • Point-in-Time (PIT) is determined by the time of session activation
• Two modes

• Copy on First Access (deferred)

• Full Copy mode

• Target device is at least as large as the source device

Pointer Based Virtual Replication:
• Targets do not hold actual data, but hold pointers to where the data is located

• Target requires only a small fraction of the size of the source volumes
• Target devices are accessible at the start of session activation
• Pointer based virtual replication uses Copy on First Write (CoFW) technology
• When a write is issued for the first time to source or target:

• Original data at that address is copied to a predefined area in the storage array
called “Save location”

• Pointers in the source/target are updated to point to “Save location”

Remote Replication: A process of creating and maintaining copies of data from a production
site to remote site(s).

• Addresses risks associated with regionally driven outages

• Modes of remote replication (based on RPO requirements)

• Synchronous

• Replica is identical to source at all times – near zero RPO

• Asynchronous

• Replica is behind the source by a finite time – finite RPO

• Network infrastructure is required between source and target
Compute-based Remote Replication:
• Replication is done by using the CPU resources of the compute system, using a
software that is running on the compute
• The remote replication methods are:

• LVM-based

• All writes to the source Volume Group are replicated to the target
Volume Group by the LVM
 • Can be in synchronous or asynchronous mode

• Database Log Shipping

• Transactions to the source database are captured in logs, which are

periodically transmitted by the source compute system to the remote
compute system

• Remote compute system applies these logs to the remote database

Storage Array Based Remote Replication:
• Performed by array operating environment

• Synchronous Replication

• Asynchronous Replication

• Disk buffered Replication (shown in the figure)

• Combination of local and remote replications

• RPO usually in the order of hours

• Low bandwidth requirements

• Extended distance solution

=============== =================

CDC Management
• Key management activities in a CDC:

• Monitoring and Alerting

• Reporting

• Availability Management

• Capacity Management

• Performance Management

• Security Management
 Monitoring: Compute systems, storage, and networks are the key components to be
monitored

Alerting of Events: Alerting is an integral part of monitoring

Monitoring tools enables administrators to assign different severity levels for different
alerts

 Reporting: Reporting on CDC resources involves keeping track and gathering

information from various components/processes

• Establishes guidelines for all configurations to achieve high availability based on

service level requirements
• Ensures high availability by:
• Eliminating single points of failure by configuring
• Two or more HBAs/NICs
• Multipathing software
• RAID protection
• Redundant Fabrics
Performing data backup and replication
 Capacity Management:
• Ensures adequate availability of resources based on their service level requirements
• Manages resource allocation
• Key activities

• Trend and Capacity analysis

• Storage provisioning

• Examples

• Compute: Compute configuration and file system/DB management

• Storage: Device configuration and LUN Masking

• SAN: Unused Ports and Zoning

 Performance Management:
• Configure/design for optimal operational efficiency
• Performance analysis

• Identify bottlenecks

• Fine tuning for performance enhancement

• Key activities:

• Compute: Volume management, database/application layout

• Storage Array: Choice of RAID type and layout of devices (LUNs) and choice
of front-end ports

• SAN: Designing sufficient ISLs with adequate bandwidth

 Security Management:
• Prevents unauthorized activities or access
• Key activities:
• Compute

• Creation of user logins and user privileges

• Storage Array

• LUN masking prevents data corruption on the storage array by

restricting compute access to a defined set of logical devices

• SAN

• Configuration of zoning to restrict unauthorized HBAs