Disaster Recovery Experts Speak Out (continued) - InformationWeek

InformationWeek is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


Disaster Recovery Experts Speak Out (continued)

THE DISCUSSION (continued)

How do you do it?
thread by Martin Garvey (02-Nov-01 2:22 PM GMT)

How is it possible to build an ironclad business continuity infrastructure if the host and backup systems are not in disparate geographic locations?

Re: How do you do it?
by Ed Willhouse (07-Nov-01 2:47 PM GMT)

Over the past 15+ years the Company downsized it's IT operations from multiple Centers to one large Center supporting the enterprise. This support model has proven to be very cost efficient. Our current DR strategy is to build at time of disaster (no cold or hot site). After many attempts I have been unable to muster support for a Hot Site strategy. Suggestions?

Re: How do you do it?
by Andreas Wolf (05-Nov-01 5:05 PM GMT)

I work for a firm that has specialized in 24x7 facilities for the past decade or so. We have put in place literally hundreds of millions of dollars in backup facilities during that time, for many clients all around the country.

The cost of large, mirror facilities is a tremendous investment, but one that has made a great deal of sense to many large corporate clients. The latest realization to hit them, though, is that they have created redundant platforms and redundant links and redundant storage, all to varying degrees of "realtime" as suits their business, but none have really created the redundancy in staff that is actually required to make efficient use of all that hardware and infrastructure. The events of September 11, and the tremendous impact of both people lost outright, and people unwilling or unable to travel to alternate locations, has highlighted this resource problem.

The hardware and software industry have long been promising that remote operations are just around the corner. To be fair, a lot of remote capability does exist today. However, there still seems to be a great deal of room for improvement, and now there seems to be an impetus to really get it to a plug and play level of capability, something even smaller enterprises can do without massive investments in technical resources or material resources.

This whole discussion and evolution leads to the following question: If human resources are one of several key driving issues in disaster recovery, does it not make more sense to avoid a concentrated data and communications environment altogether? Is it the best thing to do to grid our connections, distribute our resources, both people and infrastructure, into pods, and to work at a capacity that allows one or more pods/nodes to fail without overloading the system? Your thoughts?

Protecting Medical Records
thread by Dr. Kenneth Buckle (05-Nov-01 11:48 AM GMT)

Any thoughts about protecting electronic medical records systems? Some medical environments are moving toward paperlessness, and every piece of medical data entered is important and being updated on a 24-hour basis. Is it feasible to do a "continuous back-up" for this type of data?

Re: Protecting Medical Records
by Jeff (06-Nov-01 11:48 PM GMT)

Have you ever looked into electronic vaulting? Check out www.evaultnw.com or www.evault.com

Re: Protecting Medical Records
by Scott Kennedy (06-Nov-01 11:57 PM GMT)

Dr. Kenneth Buckle,

Your question about electronic medical records systems and the feasible to do a "continuous back-up" is a starting point for data protection. First let me answer your question, yes, a continuous back-up is feasible.

Now let me describe some of the tradeoffs and different approaches to storage management. Data can be replicated from one storage device to another storage device and allows you to have a mirrored copy of your medical records in real time, should the primary disk fail.

Data replication comes in many flavors; Host based software, hardware based storage control units, virtualization, etc. Within these different techniques are still more considerations like local mirroring (synchronous), remote mirroring (asynchronous), point-in-time copies, snapshot copies, etc. These are various ways to have a second copy of data on disk and allow you to recover the data very quickly.

In addition, for disaster recovery purposes, you must have the data archived (to tape) and located onsite and offsite (vaulted). However, there are technology and processes involved in all these approaches that can be very difficult to manage and can be very expensive.

Bottom line. There are ways to optimize your current storage infrastructure and simplify your processes without breaking the bank.

Scott Kennedy
Vice President Business Development Storage Solutions Fujitsu Softek
Telephone: (408) 992-2588
Email: [email protected]

Re: Protecting Medical Records
by Rob Young (08-Nov-01 3:09 AM GMT)

Dr. Ken,

I work in the Healthcare industry and DR has been and will continue to be a big issue. With multiple platforms we are working with one of the top 2 DR services to ensure we can recover. Some of the better solutions are actually technology based. Let me explain. One of our key systems runs on a VMS cluster. Active servers (not backup) are split across buildings as is storage. Each transaction hits disks in separate buildings. The distance separating the buildings doesn't qualify this as Disaster Tolerant but it is fault tolerant. The magnitude of the disaster to render both datacenters "down" would allow a lazy recovery to take place (i.e. would be the least of our concerns).

VMS clusters were very instrumental for World Trade Center folks to continue to carry on business... here is a link to an article that shows how Cantor Fitzgerald actually continued business having lost one datacenter. Not failover computing, but active-active computing... and VMS was/is at the heart of their operations: http://www.eweek.com/article/0,3658,s%253D701%2526a%253D17239,00.asp "When one corner [World Trade Center] went down, we never stopped trading," Kiewel said. "The system kept functioning in Europe and Asia."

Cantor Fitzgerald has a Disaster Tolerant configuration. Part of their solution is software based...i.e. they may be using RTR to ensure transaction integrity. RTR is at the heart of most financial industry backends. Financials can't afford to lose a single transaction. Here is a link to the RTR homepage: http://www.compaq.com/products/software/ntenterprise/rtr/

I realize this may not be the answer you were looking for. But the key to Disaster Recovery is to avoid it at all costs with Disaster Tolerant solutions (assuming you can afford to and even if you can't, fault tolerant may be a good fit , and quite affordable actually).

Oh... what would I say is fault tolerant versus Disaster Tolerant? A rule of thumb would be to have datacenters at least 10 miles apart on separate power grids. This is more costly infrastructure wise. I can use switched in-house network connections between datacenters. At 10 miles, you are doing ATM with multiple leased lines from different carriers to ensure (hopefully, but there are horror stories there too...) connections between datacenters.

Finally, one reason to trumpet DT versus DR. The DR drills are scary for the most part. Scary in the sense, if they went well , I'm shocked to hear that (and I have heard my share of those stories too :-)

Rob Young VMS Consultant

Re: Protecting Medical Records
by Jason Buffington (08-Nov-01 3:45 AM GMT)

YES - it is possible to protect that kind of data. The real trick (without sounding like a commercial for the product that I represent) is to use a technology that replicates "file systems" and not "applications". The difference is that instead of plugging into SQL or Oracle or some other top-tier application - file system replication protects any file on the supported Operating System. And most (including mine, obviously) on a real-time basis. So, as soon as new data is written to the production server, it is also written to a secondary platform.

"Continuous Backup" means different things to different people. For most people, it translates to "real time replication" or "always protected". For those that really mean "Continuous Backup" as routinely writing to a tape device, almost every backup technology can be leveraged against a secondary copy of data (provided by replication software).

So now, a replication technology ensures that if the whole production server dies, an always available and current copy exists. Additionally, backup technology can capture only those files that have changed in the last two or four hours (from the replicated copy of storage) and give you images of any file, at any granularity. This is key for software developers and others who need multiple iterations of any file at any time.

TIP: The solution starts with replication software, so that granular backups can be done without impacting the production server(s).

Re: Protecting Medical Records
by Eddie Anderson (08-Nov-01 5:14 AM GMT)

Dr. Buckle:

I have been in healthcare IT my entire career and my wife is IT director for a 50 something provider multispecialty group. Unless you have the megabucks of only a few top medical centers, disk mirroring and nightly back-up will give you enough protection that you should sleep very well knowing that what you are doing is much more fault tolerant than paper or single disk and more secure, if your network is designed properly. I always recommend disk mirroring and, using my wife's practice as an example, it has proved worthwhile twice in the past seven or eight years. With today's cost of storage it is also very economical. It will also go a long way toward helping in your HIPAA compliance (not security/privacy but the availability/disaster recovery piece. No, it is not "disaster proof" but neither HIPAA nor "reason" dictate that and 99.9% of healthcare can't afford it.

Re: Protecting Medical Records
By Adam Garsha (08-Nov-01 11:44 AM GMT)

In some respects, "continuous back-up" is the norm. EMR software along with all other forms of client/database server software can most often make use of something called 'journaling.'

Essentially, every change that you submit to the database; e.g. little Jimmy has X and X symptoms, I'd like to place the following orders, etc., etc. will be recorded not only to the database, but to a separate journal file as well.

Journal files along with nightly backups (to tape or other media) provide the ability to fully recover a database system with up-to-the-minute accuracy.

Of course, the downtime associated with a catastrophic restore is a nightmare. Consequently, significant hardware redundancy and HA/clustering is employed to ensure that... you never need to go to backup.

Re: Protecting Medical Records
By Patrick Fern (08-Nov-01 3:38 PM GMT)

While the mindset today "I want a DR strategy, because I know I'm going to need it" is a valid one, it is possible to tolerate interruptions in service as opposed to picking up the pieces later.

Disaster tolerance as you describe exists on the Windows 2000 and NT platforms as well. Marathon Technologies uses Split-Site (tm) to maintain the uptime of the application by geographically separating the two halves of the server array by up to 10K. There is no single point of failure. This will tolerate a disaster event, even to the level of total destruction of one half of the array, while the application continues through with no failover latency or split-brain problems. End users are not required to re-connect, in fact, they don't even notice the disruption in service.

Off-the-shelf or custom software runs with no modifications and no scripting, and the use of standard server hardware is permitted as well.

Patrick Fern
Customer Relationship Mgr
Marathon Technologies Corp.
978-489-1152 [email protected]

Re: Protecting Medical Records
by Rob Young (08-Nov-01 6:30 PM GMT)

"File system replication" is a product that will shrink into its niche. Future storage solutions for Compaq include DRM II. I have read that " DRM II has the ability to access DRM-mirrored units from both ends simultaneously" depending on host capability... at the very least keeping sites in lockstep, transparent to OSes (i.e. at the block level). For OSes that support lock managers, the advantage with DRM II will be the ability of reading from either site at the same time. Active-active instead of "hot standby". Quite annoying to have remote disks that are doing little more than catching writes. I want to be able to read from them. After all, most folks DBs are doing 80/20 Read to write ratios at the disk level (maybe 60/40 if you really have your caching act together).


<< Previous Page | Next Page >>

We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.
Comment  | 
Print  | 
More Insights
2021 State of ITOps and SecOps Report
2021 State of ITOps and SecOps Report
This new report from InformationWeek explores what we've learned over the past year, critical trends around ITOps and SecOps, and where leaders are focusing their time and efforts to support a growing digital economy. Download it today!
InformationWeek Is Getting an Upgrade!

Find out more about our plans to improve the look, functionality, and performance of the InformationWeek site in the coming months.

Remote Work Tops SF, NYC for Most High-Paying Job Openings
Jessica Davis, Senior Editor, Enterprise Apps,  7/20/2021
Blockchain Gets Real Across Industries
Lisa Morgan, Freelance Writer,  7/22/2021
Seeking a Competitive Edge vs. Chasing Savings in the Cloud
Joao-Pierre S. Ruth, Senior Writer,  7/19/2021
Register for InformationWeek Newsletters
Current Issue
Monitoring Critical Cloud Workloads Report
In this report, our experts will discuss how to advance your ability to monitor critical workloads as they move about the various cloud platforms in your company.
White Papers
Twitter Feed
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.
Sponsored Video
Flash Poll