The California Energy Crisis “Illuminates” Cutting Edge Data Protection and Availability Solutions

– by Scott Robinson

While regulators and public utility companies grapple with the economic issues of the California energy crisis, businesses are left scrambling, wondering if their information systems are prepared for the inevitable power outage. The energy crisis in California has prompted utility firms to institute power outages to save energy, affecting millions of customers statewide. With a crisis like this comes the often evaded question: “Does our current information infrastructure contain the most sophisticated data protection technologies including backup and recovery, mirroring, and power protection, and is it enough to maintain 100% availability should this crisis or any other disaster occur?”

Data Protection and Availability Solutions
There are many data protection and availability options available ranging from simple CD-RW discs for backing up PC hard drives to full-scale Storage Area Networks (SANs) based systems that allow data backups to be performed and data mirrors to be created across an enterprise-wide shared server/storage environment. Standard technologies used industry-wide for data protection and availability include UPS devices and portable generators to safeguard against power failures, remote backup and archiving, enterprise remote copy and data mirroring, and high availability/clustering solutions. The following summarizes each technology and how they work to provide business continuity.

Power Protection Solutions:
A UPS device offers protection against all types of power fluctuations and failures. Online UPS devices are able to supply continuous “clean” and stable power by acting as a bridge (middle-man) between your equipment and the AC line power from the utility company. The UPS’ battery is continually recharged as long as normal line-power is being supplied. When a power disruption or power out-of-tolerance condition occurs, the battery supplies enough power to run your equipment for a specified duration. If the system must eventually be shut down due to a lack of power, the UPS, working in conjunction with the operating system’s emergency procedures, will safely provide power to store all active data and enable an orderly system (or network) shutdown. There is no transfer time needed to switch from standard utility power to the UPS supplied power since the device is continually online.
For extended power outages, battery based UPS systems may not be adequate. Today, enterprises with critical data needs augment their UPS systems with portable power generation equipment. Small portable power generators run on gasoline. Larger units often run on natural gas or propane, and can provide power as long as fuel is fed into them. Though not recommended to bring an entire site up to full-power, a generator can keep essential communications and computer systems running and operable for extended periods of time. Portable generators can be configured to supply power automatically and transparently during extended outages.
UPS and power generators protect against power disruptions. They do NOT protect your data. In the case of natural disasters, hackers, viruses, employee sabotage, or system or telecommunications failures, having recently archived or duplicate copies of all critical data is the key to getting back to “business as usual”. To be safe against natural or man made disasters which can prevent physical or electronic access to the building or network, those backup copies must be stored off-site, either at a self-owned facility, with a co-location partner or with a disaster recovery vendor. Today, access to this remote data is often enabled using wide area network (WAN) connectivity. The solutions showcased below are the cutting edge technologies being implemented for off-site data storage today.

Backup and Recovery Solutions:
Still the most fundamental and cost effective way to protect data, backup and recovery systems have evolved dramatically from the days of stand-alone tape drives contained within each server. Automated backup and recovery of data using centralized, automated tape library systems has become pervasive in all but the smallest environments. These systems can now take advantage of the most recent networking and data management software technology.
Gigabit Ethernet: Standard Ethernet networks now support speeds up to one gigabit (Gb)/second. Dedicated LAN segments at Gb speeds are often adequate for backup and recovery traffic to a centralized backup server and tape library.
SANs and Fibre Channel: A Storage Area Network (SAN) is a storage network architecture that provides an interface between multiple servers and mixed storage devices. The SAN acts as a type of storage repository, moving data storage off the LAN onto a subsystem that is usually connected to the LAN by way of a Fibre Channel fabric. Fibre Channel (FC) supports high-speed communications between any server and any storage unit, and also enables direct storage-to-storage data transfers.
For large, high-performance environments, Fibre Channel has become the storage network of choice. With switched Gbit speeds (100 Mbytes/second) soon moving to 2 Gbit/second, Fibre Channel offers plenty of storage tuned bandwidth. Fibre Channel also enables high performance drives to be directly allocated to file and application servers for optimum performance. These drives can also be dynamically re-allocated to other servers for efficient use of tape drive resources.
Besides providing unsurpassed data transfer speeds of up to 100 MBps (moving to 200 MB/s), FC has the ability to increase the distance between the main network and the storage devices up to a maximum of 6 miles (10KM). This allows data to be copied to an off-site storage device such as a tape library or RAID unit located several miles from the worksite. Data stored off-site is protected, as it is not subject to disasters that affect or prevent access to the main facility.

Server-free Backup: The next step in the evolution of backup and recovery architecture is the advent of “server-free” backup. This technique allows the movement of backup data directly from disk to tape, eliminating the involvement of the application server in the CPU intensive step of data transfer. This becomes especially powerful when combined with snapshot capability to transparently and regularly create point-in-time copies of live system data.
With server-free backup, applications remain on-line 24X7 and are always available while data protection is maintained. Data backup snapshots are continually taken without the need for backup windows. In a SAN environment, disks attached to large, business-critical servers can be backed up directly to tape, under the control of a central device, taking advantage of SAN topology.
Remote Connectivity: Other recent advancements in the area of SAN connectivity will enable remote backup and archiving over much longer distances. Channel extenders are components that allow extended distance data transfer between storage devices across separate Fibre Channel SAN islands. For example, these extenders can be attached to Fibre Channel switches in geographically separate locations, allowing SAN connectivity over WAN infrastructure, enabling remote backup of critical system data. Channel extenders greatly expand the geographical boundaries for data storage making efficient and timely remote backup and disaster recovery over long distances a reality.

Enterprise Remote Copy and Data Mirroring Solutions:
Replication software solutions provide for seamless wide area data availability and disaster recovery. For example, data can be mirrored between remote locations over any IP network for maximum business continuity. This data replication provides a storage independent solution to deliver true disaster recovery when data currency and availability are paramount.
Replication software can also ensure that multiple file servers are fully synchronized for maximum availability and load balancing. Replicated files (whether local or remote) are always available for use at secondary systems. Fully synchronous replication ensures that all participating nodes are up-to-date at all times. Some replication software offer true bi-directional replication; for example, a two-node configuration supports both servers reading and writing to the replicated data set with complete integrity.
All remote copying and data mirroring solutions ensure continuous access to the stored files by creating multiple copies on alternate systems in real time. Because the copying process is on-going and shadows normal operations, this provides the most transparent availability of data when disasters strike.

High Availability/Clustering Solutions:
To approach 100% availability, we can move beyond protected storage and introduce redundancy at the server level. Clustering software is available for all major operating environments, and creates an environment that can withstand the loss of servers, as well as storage.
Clustering software monitors and controls the availability of applications running in a cluster. It restarts applications in response to a variety of hardware or software faults.

Conclusion
A complete data protection and disaster recovery plan requires the seamless integration of a number of sophisticated technologies. Fortunately, there are vendors who specialize in the technical analysis, integration, and implementation of backup and disaster recovery solutions into existing infrastructures.
Advanced data storage technologies are continually evolving, providing a broader, ever-growing range of options for effective protection and availability of data. Storage management specialists can match your current IT environment with the appropriate state-of-the-art solution. Don’t wait until you get caught in a California-type energy crisis, or some other kind of business-risking disaster. Take steps to incorporate the new, cutting edge data protection and availability technology now available. If that disaster does strike, your business will keep on running without missing a beat.

Scott D. Robinson is the Chief Technology Officer for Datalink. He joined the company in 1989 as its Chief Engineer and became Vice President - Engineering in 1993. He was named Chief Technology Officer in 1999. Between 1983 and 1989, he was employed by Minnesota Mining and Manufacturing Company, most recently as an Advanced Electrical Engineer in the Digital Imaging Applications Center.

©Copyright 2001 Systems Support Inc. All rights reserved. Reproduction in whole or in part in any form or medium without the express written permission of System Support Inc. is prohibited.

«BACK to the Articles Index