Introduction to ZSL and Practice Area Backup and Recovery - Terms and Definitions Identifying and Defining Recovery Objectives Trends and Best PracticesQ &A
15+ years Global Technology Integrator & Business Solutions Provider, Headquartered in Edison, NJ State-of-the-art Technology Research & Development Centers in US, Canada and India 4000 employees with offices in US, UK, Canada, France, Germany, Malaysia, Singapore, Middle-East and India Dedicated R&D Division to Offer Value Added R&D Services & Product Development Services to the ISVs and SPs Emerging Technologies Specialization with the leading technology vendors alignment Pioneer in Industry Solutions Development (Insurance, Finance, E-Governance, Consumer Electronics, Pharmaceutical & Telecom) Award Winning & Proven Partnership Program “Get IT Together” Partnership for ISVs, VARs, SPs and SIs ISO and CMM Certified Solution Provider
Enterprise Infrastructure and Managed Services Based out of Edison, NJ and Chennai, India. IDEA Lab – Value-added R&D, Product Development/Engineering Portfolio of Technologies and Professional Services: Virtualization – Server Virtualization, Virtual Desktop Infrastructure, Private Cloud Implementations Data Center and Unix/Linux/Windows Services – Assessments, Implementations, Deployments and Migrations Managed Services – Manage, administer and maintain servers, networks, IT Infrastructure, Backup/Recovery services.
Recovery Time Objective (RTO) – Maximum elapsed time required to complete recovery of application data.Recovery Point Objective (RPO) – Point in time to which application data must be recovered to resume business transactions.Continuous Data Protection (CDP) – A highly dependable means of real-time, continuous data replication. Required for aggressive RTO/RPO.Replication – Anything (software or hardware) that can make additional copies of existing data to improve data availability.Virtual Tape Libraries (VTL) – Disk-based backup medium which simulates tape medium.D2D/D2D2T – Disk to Disk; Disk to Disk to Tape.
Snapshots/Copy on Write/Point in Time Copy – A snapshot is a set of reference markers, or pointers, to data stored on a disk drive, on a tape, or in a storage area network.
De-duplication – A method of reducing storage needs by eliminating redundant data.
Combined Responsibility – Recovery objectives should be defined by data/application requirements, business policies and IT processes. Define RPO/RTO – This will determine the type of backups, frequency, medium, geographical and level of redundancy. The medium is the message – Will you be able to restore from the chosen medium in 5/7/10 years time? What will be the cost for long term storage? Length of Retention – Regulatory, compliance and policy driven. Business Continuity and Disaster Recovery Plans – Backup and recovery should be part of the overall BCP/DRP. BCP/DRP starts here and goes much beyond.
Disk-based Backups – Data is backed up from high-performance disks to low- cost commodity disk systems. Combined with deduplication, snapshots and compression, disk-based backups allow for very fast recoveries and return to service. VTLs are being replaced by multi-tiered D2D and D2D2T technologies. By 2014, advanced disk-based backup will capture 80% of the overall $5-6 billion market (Gartner, January 2011). SAN to SAN Replication – Data is replicated from primary site/server to secondary site/server. Built-in SAN feature or third-party software built around virtualization technologies. Heterogeneous replication is big advantage of using software based replication. Snapshots – Virtualization is driving snapshot deployment and adaption. Makes for near-instantaneous backups and aggressive RPOs. Deduplication is maturing from a stand-alone to becoming an “must have” feature.
Backing up to the Cloud – Data is backed to an “cloud” based offsite facility. You pay for the space used on a monthly basis. Typically being used for “last tier” and archival data. Typically data is encrypted during transit and at rest. Gartner: 70% of all businesses will consider Cloud- based backups in the next 12-18 months. Continuous Data Protection – CDP (and near-CDP) is an increasingly affordable and required method of data replication. CDP-based solutions can provide fine granularities of restorable objects ranging from crash-consistent images to logical objects such as files, mail boxes, messages, and database files and logs “Bare-Metal” Recoveries – Increasing use of virtualization technologies is changing the meaning of “bare-metal” recoveries.
Backup to S3 – Run a traditional backup utility and dump the data to Simple Storage Service (S3) which is inexpensive and is available in very large quantities. Make sure the backups are on redundant availability zone and regions. RAID built from S3 – Use S3-based virtual disks to build your own RAID to mirror and replicate your data onto multiple S3 volumes. Data protection comes from RAID’s high-availability and mirroring. Virtual SAN/NAS – Build a virtual iSCSI SAN or NFS/CIFS NAS in EC2 so that multiple instances can served and backed up from a single instance. Replicated/Distributed FS - Use a distributed filesystem like GlusterFS so that your EC2 instances can replicate each other’s data. Make sure there are a large number of participant instances since this increases the redundancy and availability of the data.
Policies and Processes – Backup and restore procedures and infrastructure should be built upon and driven by business policies and IT processes. Define Recovery Objectives First – Recovery objectives should drive backup decisions and procedures, not the other way around. For example, data backed up to the Cloud can take a long time to recover; so this is not ideal for applications that have an aggressive RTO. Test Regularly and Randomly – A backup that cannot be restored is a waste of effort, time and resources, besides building a false sense of security. Test and verify your backed up data and recovery procedures at regular intervals. Mix in some randomly times fire-drills to check skills and training levels. Prioritize and Classify – Work with all the stake holders to define and prioritize your data into different tiers based on business importance and policies. Build your backup/recovery infrastructure accordingly. Re-visit these decisions at regular intervals. Think Long Term – Data sticks around for a long time. Chose your backup methods and media such that data is recoverable at the far-end of the retention periods. Consider costs for the long term as well. For example: It costs a lot to keep disks spinning for 5 to 7 years; tapes can be store data without burning watts. Make sure you complete an ROI analysis before committing.