From the year 2000 onwards, 64-bit servers will be standard, but will companies be able to utilise the power that this new platform will provide?
Lewistown Hospital in central Pennsylvania is a 228-bed, non profit community facility with approximately 8,400 admissions and over 132,600 outpatient visits per year. In December of 1997, the hospital replaced its OpenVMS VAX equipment with a 64-bit DIGITAL AlphaServer OpenVMS cluster configuration running SMS ALLEGRA 2.3.
According to Ron Cowan, director of Management Information Systems, the move to clustered computing has been a great success. "The cluster delivers significant benefits from both an availability and a productivity standpoint," he says.
"We run our production processes on one node and our file conversions on the other. This means we can move ahead with the important conversion process while keeping our online user community happy. At the same time, we have the knowledge that the cluster failover capabilities can keep us up and running in the event of a problem."
In addition to performance and availability, Cowan cites the Alpha systems rackmount configurations as another plus. "The small footprint of the rackmount cabinets is a big advantage in the tight quarters of the hospital data center."
Lewistown Hospital has been an SMS ALLEGRA shop since 1989. With the recent upgrade to ALLEGRA 2.3, the hospital runs the entire ALLEGRA suite of applications including Clinical Data Software (CDS), Patient Data Software (PDS), and Administrative Data Software (ADS). The hospital also runs SMS Pharmacy and SMS Radiology on the powerful cluster configuration.
The cluster is composed of two AlphaServer 4100 systems, each with two CPUs and two Gb of RAM running the OpenVMS operating system and OpenVMS Cluster software. The cluster was factory preconfigured by DIGITAL
CustomSystems Value-added Implementation Services (VIS) significantly reducing installation time and operational disruption.
Lewistown Hospital systems analyst Lori Cottrill reports that the cluster has delivered spectacular improvements in performance. "We run nursing worksheets three times a day," explains Cottrill. "The program logs and tracks patients' locations and medical orders, and prints out a worksheet at the nursing stations in time for each shift change. In the past, the program took forty minutes to run. Today, on the cluster, it takes a minute and a half. An end-of-month report that uses DATATRIEVE to generate internal reports from ALLEGRA data files finishes in forty minutes, compared with 16 hours in the past. And on the financial side of the house," reports Cottrill, "An electronic remittance program that formerly took two hours to run now executes in 15 minutes."
Lewistown Hospital is using software provided by SMS to convert hundreds of thousands of patient files ( seven years of historical data ( into the new ALLEGRA file format. When a file is converted it becomes part of the ALLEGRA database. Because the conversion is being done on the cluster ( a virtual single system ( the conversion is rapid and seamless.
Bob Bannon, MIS operations supervisor, reports that the cluster supports a network with 150 PCs, while an additional 150 people connect to the system with terminals. There are also 30 network-connected laser printers. At peak usage, the cluster handles 290 processes on one node.
Bannon says that with dual HSC50 storage controllers "our I/O capabilities are extremely high". He has implemented OpenVMS Volume Shadowing on the 64 disk drives providing 120 Gb of online storage. He backs up the online storage using TZ877 DLT drives with autoloaders enabling a week's worth of backups.
While one cluster node takes care of backup functions and runs the conversion process, the other node is geared exclusively towards the clinical user community. "One of the first responses we received after the new systems were installed was from the radiologists. They were delighted by the speed differences, especially in the refresh rate of interactive screen displays," notes Lori Cottrill.
The ultimate configuration
Ron Cowan says: "The cluster is the ultimate configuration. MIS gets maximum productivity, users get rapid response, and we have the failover capability should we need it.
"We are very impressed with the failover capability," concludes Cowan. "We ran it in the testing phase and it performed perfectly. It's like an insurance policy. We hope we never need it but it's good to know it's there."
Compiled by Paul Phillips
(c) Compaq 1999
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Read more on IT risk management