BACKUP/MASTER: Strategies for Archiving Dianne McAdam Senior Analyst and Partner Data Mobility Group
Backup versus archiving Terms used interchangeably Define different processes Different end results
Backups Designed to protect ALL data Datacenter, desktop, laptops Run on a regular basis Every 24 hours More often for mission-critical applications Multiple backup copies are saved Several copies of weekly backups Incrementals Monthly and quarterly backups
Backups (2) Usually not actively deleted Overwritten Second set of incrementals overwrites the first Fifth set of weekly backups overwrites the first Multiple backup processes and software can exist One product for laptops Another for departmental backups
Archive Not a backup copy Not a disaster recovery copy Data that is kept for a long period of time Can be called: Fixed-content data Unstructured data Reference data Retention managed data
Archive (2) Active archive Deep archive
Active archive Software to “slim down” databases Databases expand over time Customer databases contain numerous “inactive” customers Software extracts inactive data Associated table definitions, indexes, metadata Migrates inactive data to lower-cost storage Disk, tape or optical storage
Active archive (2) Designed to improve management of large databases Not a replacement for backup Not a form of data protection Does not have to run daily, but on a regular basis Weekly, monthly Depends on the amount of data that becomes “inactive” Saves money Using lower-cost storage Improves performance of main database Quicker to backup and restore
Deep archiving Corporations have had internal policies for how long data should be kept Medical, financial, employee records Engineering designs New regulations are dictating retention periods Archival software designed to store regulated and non-regulated data for specific period of time
Deep archiving (2) Not a backup copy Not a disaster recovery copy Data is actively deleted When a specified time period elapses Seven years for When a specific event occurs Two years after an employee leaves the company
Deep archiving (3) Many different types of storage Tape, optical, disk, special appliances Backups store multiple copies of data; archive should store only one copy SEC 17a-4 Need to keep one copy offsite Some regulations require WORM support
Deep archiving (4) Recovery Time Objectives (RTO) Time to recover data Different from backups Recovery Point Objectives (RPO) Age of data used to restore RPO = 0
What’s the cost? Not just acquisition costs Other long-term costs Environmental factors Maintenance
Hardware requirements How much data will be archived? Projected growth rate Retention period Backed up onsite or offsite Extra capacity for unexpected growth
Hardware requirements (2) Media choice Disk, SATA disk, tape, optical Average disk/tape utilization RAID level Combination of disk/tape? WORM support required?
Purchase and maintenance costs Purchase price Installation and freight charges Warranty period Lifespan of equipment Cost to migrate People Software
Software requirements Archive software costs Additional software required? Compatibility with existing/new equipment
Environmental requirements Footprint Power Cooling
Calculating the costs Calculate storage requirements Assumptions Initial storage requirements = 50 TB Annual growth rate = 10% Average tape utilization = 85% Average compression 2:1 Average disk utilization = 70% No data expires the first 7 years
Calculating the costs Terabytes Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Initial Storage New Data Total Data Disk Requirements Tape Requirements
All-tape configuration Tape Library$77,860 4 LTO2 Drives$73, Cartridges$18,655 Total$170,075
All-disk configuration 10 controllers$243, expansion units$695,490 4 racks$18,000 Total$957,300
Calculating environmentals Assumptions $20/square foot/month $.10/KWH Electrical = (Power + cooling) * units * 8760 *.1 Footprint Yearly Cost Power per unit Cooling per unit Electrical per year Total Library10.2$ KWH $1612 Drives0$0.032 KWH $ 224 Total$1836$4284
Calculating environmentals (2) Assumptions $20/square foot/month $.10/KWH Electrical = (Power + cooling) * units * 8760 *.1 Footprint Yearly Cost Power per unit Cooling per unit Electrical per year Total Disk Base unit 29$6, KWH.39KWH$6,307 Expansion unit 0$0.33 KWH.39 KWH$18,921 Total$25,228$32,168
Seven-year costs Year 1Year 2Year 3Year 4Year 5Year 6Year 7Total Tape$174,359$4,284 $200,063 Disk$989,468$32,168 $1,182,476
Some caveats – not included The cost of maintenance The cost to replace disk and tape The cost of the software The cost of phasing in equipment purchases
Some caveats – not included (2) The amount of data that will expire The cost associated with backing up data The additional cost for RAID-1 or RAID-5 disk The additional cost of HBAs and cables, etc.
Other options Hybrid disk/tape solution RawUsable Disk Capacity Tape Capacity Total12098
Hardware costs - hybrid Disk appliance $276,600 Tape Library $76,640 2 LTO2 drives $36, cartridges $15,470 Total$128,890
Seven-year costs: Hybrid solution Year 1Year 2Year 3Year 4Year 5Year 6Year 7Total Tape$133,062$4,172 $158,094 Disk$284,285$7,685 $330,395 Total$488,489
Summary Disk Initially costs more and continues to cost more over the years Faster retrieval time How often do you replace disk? Every 3 years? Tape Initially costs less and continues to cost less Slower retrieval time than disk How often do you replace tape? Every 5 – 7 years?
Summary (2) WORM support WORM cartridges cost 10-15% more Evaluate tape/disk combination Evaluate integrated tape with disk appliance Evaluate migration efforts and costs