Virtualizing Sectra RIS/PACS Solutions for Healthcare ... · Virtualizing Sectra RIS/PACS Solutions...

Virtualizing Sectra RIS/PACS Solutions for Healthcare Providers

Enabled by EMC Unified Storage and VMware Applied Technology

EMC Global Solutions

Abstract

This white paper provides an overview of a Sectra RIS/PACS environment built on VMware® ESX® with EMC® Celerra® and CLARiiON® storage. It describes the tested configuration environment and identifies the key results of testing. This white paper covers EMC Unified Storage and VMware vSphere™.

December 2010

Virtualizing Sectra RIS/PACS Solutions for Healthcare Providers Enabled by EMC Unified Storage and VMware 2

Copyright © 2010 EMC Corporation. All rights reserved.

EMC believes the information in this publication is accurate as of its publication date. The information is subject to change without notice.

THE INFORMATION IN THIS PUBLICATION IS PROVIDED “AS IS.” EMC CORPORATION MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE INFORMATION IN THIS PUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Use, copying, and distribution of any EMC software described in this publication requires an applicable software license.

For the most up-to-date listing of EMC product names, see EMC Corporation Trademarks on EMC.com

All other trademarks used herein are the property of their respective owners.

Part Number h8096


Table of Contents Executive summary ............................................................................................ 4 Introduction ......................................................................................................... 4

Audience ...................................................................................................................................... 4 Terminology ................................................................................................................................. 5

Environment overview ........................................................................................ 6 Architecture overview ................................................................................................................... 6

Architecture diagram ................................................................................................................ 6 Test environment overview .......................................................................................................... 7

VMware vSphere ...................................................................................................................... 7 Celerra NS-960 / NS-120 ......................................................................................................... 8 Celerra Replicator .................................................................................................................... 8 LUN migration........................................................................................................................... 8 EMC MirrorView/S .................................................................................................................... 9 EMC Unisphere Analyzer ......................................................................................................... 9 Enterprise Flash Drives ............................................................................................................ 9 Celerra FLR – NAS file systems ............................................................................................... 9 Sectra RIS / PACS ................................................................................................................... 9

Test results ........................................................................................................ 11 VMware HA ............................................................................................................................ 11 VMware vMotion ..................................................................................................................... 11 VMware DRS .......................................................................................................................... 11 LUN migration......................................................................................................................... 12 Celerra NAS share for archiving............................................................................................. 13 Celerra Replicator .................................................................................................................. 14 EMC MirrorView/S .................................................................................................................. 14 Celerra File Level Retention ................................................................................................... 14 iSCSI ...................................................................................................................................... 14

Conclusion ........................................................................................................ 15 References ........................................................................................................ 16


Executive summary Healthcare organizations around the world continue to invest in radiology and cardiology Picture Archiving and Communication System (PACS) to predict, diagnose, treat, and monitor disease. With each new imaging modality technology advancement, healthcare provider IT organizations are challenged to meet increasing demands for high availability of systems and networks, security, and automated tools to simplify the complexity of managing this expanding imaging environment.

Healthcare providers have also recognized the value of virtualization in helping reduce the cost and complexity of their IT operations and many have begun deploying virtualized IT environments to support their PACS and other mission-critical healthcare applications. This has resulted in higher expectations from the IT infrastructure to provide high availability, security, and tools to support these clinical applications.

EMC offers several key infrastructure components to economically deliver highly available applications to healthcare IT consumers. These components, when combined with clinical and business application software from EMC partners, deliver continuous information availability and protection to enable continuous hospital operations. Some of these components include:

• Tiered-networked storage platforms that provide the availability, reliability, scalability, and performance needed by the healthcare enterprise.

• Server virtualization to dynamically map computing resources to the healthcare enterprise. This results in a lowering of IT costs due to the ability to treat the data center as single pool of processing.

• Information management and protection software to meet recovery point and recovery time objectives for business continuity and disaster recovery plans, and to maintain the privacy and security of patient records, confidential e-mails, and other communications.

• Active archiving components to provide secure and rapid recall of historical images and to meet regulatory requirements for records retention.

This information infrastructure, built for reliability, supports both short-term access and long-term archiving and can be integrated with cardiology, radiology, and other clinical and business systems to automate workflow and streamline operations.

Introduction This white paper provides an overview of a Sectra RIS/PACS environment built on VMware® ESX® with EMC® Celerra® and CLARiiON® storage. It describes the tested configuration environment and identifies the key results of testing. The white paper includes the following sections: • Environment overview • Test results

Audience This document is intended for stakeholders including the EMC Global Practice, Product Engineering, Field Council, Information Infrastructure Solutions Engineering, Marketing, and Demo teams as well as EMC Training and Services organization.


Terminology EPR Electronic Patient Record: An information system for keeping records of the periodic care provided by an

institution to a patient. HIS Hospital Information System: Typically an integration of order-entry systems, an administrative system,

and departmental subsystems within a hospital or hospital chain. NAS Network-Attached Storage: Typically a one-box solution with hard-disk storage bundled with a server

that is set up with its own network address. PACS Picture Archiving and Communication System RIS Radiology Information System SAN Storage Area Network: A high-speed special-purpose network connecting various sorts of storage devices

with servers, typically to support a larger network of users.


Environment overview

Architecture overview The following section identifies and briefly describes the technology and components used in the validated virtualized desktop infrastructure environment.

Architecture diagram The following illustration shows the architecture of the virtual desktop infrastructure discussed in this white paper.

Figure 1: Architecture diagram


Test environment overview The storage configuration, see Figure 1, tested consisted of an EMC Celerra NS-960 unified storage platform with FC connectivity between three Dell R710 servers with ESX 4.1 (4.1.0-0.0.260247) installed. The fabric consisted of two Brocade 4100 switches. The network switch was a Cisco Catalyst 3560G.

The native FC ports of the NS-960 were connected to the SAN. This allowed the back end of the NS-960 to present FC LUNs to the virtual machines as raw disk mappings (RDMs) through the ESX servers.

The Celerra presented CIFS shares to the virtual machines, and these shares were used for archiving the data from the ImageServer. Celerra Replicator™ was configured for this file system between the NS-960 and the NS-120.

EMC MirrorView™/S was configured between the back-end NS-960 and the NS-120 for the storage that was presented to the ImageServer.

VMware vSphere VMware vSphere™ 4.1 is a highly reliable platform for data center virtualization. It enables scalable and efficient use of x64 server hardware in a robust, highly available environment. The VMware ESX server:

• Abstracts server processor, memory, storage, and networking resources into multiple virtual machines, forming the foundation of the VMware vSphere 4.1 suite

• Partitions physical servers into multiple virtual machines, each representing a complete system with processors, memory, storage, and BIOS

• Shares single server resources across multiple virtual machines and clusters ESX servers for further sharing of resources

vMotion VMware vMotion® provides the capability to move a running virtual machine from one ESX/ESXi host to another. It eliminates the need to schedule application downtime due to scheduled server maintenance through live migration of virtual machines across servers with no disruption to users or loss of service. Storage vMotion VMware Storage vMotion provides the capability to move a running virtual machine from one storage device to another. It eliminates the need to schedule application downtime due to planned storage maintenance or during storage migrations by enabling live migration of virtual machine disks with no disruption to users or loss of service. Distributed Resource Scheduler (DRS) VMware DRS works with vMotion to provide automated resource optimization and virtual machine placement and migration, to help align available resources with predefined business priorities while maximizing hardware utilization. High Availability (HA) VMware HA provides easy-to-use, cost-effective, high availability for applications running in virtual machines. In a server failure, the affected virtual machines are automatically restarted on other production servers with spare capacity.


Celerra NS-960 / NS-120 EMC Celerra unified storage system technology is based on breakthrough architecture and extensive technological innovation, providing consolidated, multi-protocol storage that delivers solutions that enable healthcare IT organizations to meet requirements from the data center to remote clinics or imaging centers. The unified storage with UltraFlexTM technology delivers industry-leading application performance, and new levels of scalability, reliability, and flexibility. Each Celerra system consists of one or more X-Blades. The X-Blades provide the interface to the IP network and are responsible for presenting the NAS shares or iSCSI LUNs.

Figure 2: Components of the Celerra unified storage architecture

Celerra Replicator Celerra Replicator (V2) enables you to create and manage replication sessions, each producing a read-only, point-in-time copy of a source object at a designated destination. Celerra Replicator (V2) sessions, which can be created by using the command line interface or Celerra Manager, are characterized by an architecture based on snapshot-checkpoint technology, asynchronous transfer to the destination, and support for file system, Virtual Data Mover, and iSCSI LUN source object types.

LUN migration LUN migration allows users to seamlessly migrate data between LUNs within an array without disruption to applications. A user can migrate data to LUNs with better performance or other characteristics within the array, providing a greater level of control over the system and higher levels of service for business applications.


EMC MirrorView/S EMC MirrorView/S offers storage system-based remote mirroring. It is designed as a disaster recovery (DR) solution for mirroring local production data to a remote/disaster recovery site. It provides end-to-end protection by replicating the contents of a primary volume to a secondary volume. MirrorView/S is a synchronous product that mirrors data in real time between local and remote storage systems. Features such as consistency groups allow applications that are write-order dependent across volumes to manage several volumes as one volume. Data replication is supported over storage area networks (SANs) and IP-extended SAN environments.

EMC Unisphere Analyzer EMC Unisphere™ Analyzer gathers storage-system performance statistics and produces charts of the data. You can use these charts to identify and anticipate bottlenecks in the disk storage component of a computer system. Analyzer can also provide information for load balancing, capacity planning, or cost-benefit calculations.

Enterprise Flash Drives Enterprise Flash Drives deliver extremely high performance, and are recommended for applications that require fast response times or very high throughput (IOPS). Flash drives are built on solid state drive (SSD) technology that has no moving parts. The absence of moving parts makes these drives highly energy-efficient, and eliminates rotational latencies. Therefore, migrating LUNs from FC to Flash can both boost performance and create significant energy savings.

Celerra FLR – NAS file systems Celerra File Level Retention (FLR) is a Celerra Network Server software feature that protects files from modification or deletion until a specified retention date. When you use FLR, you can archive data to file-level retention storage on standard rewriteable magnetic disks through NFS or CIFS operations. FLR enables you to create a permanent, unalterable set of files and directories, and ensures the integrity of data.

Sectra RIS / PACS The steadily increasing size of image stacks is generally agreed to be one of the most important issues of the future for radiology. The development is driven by the desire to improve diagnostics and the quality of care. And it is clear that the significant increase in image production will continue. This entails efficiency problems in the flow of radiological work, from both a technical and human perspective. Radiology Information System (RIS) Sectra RIS is a workflow engine for paper-free radiology departments. It integrates PACS, Hospital Information Systems (HIS), and Electronic Patient Record (EPR). Sectra RIS increases efficiency by coordinating all components of enterprise workflow, including planning, Admit Discharge Transfer (ADT), order-entry, scheduling, optimal patient throughput and documentation, reporting, follow-up, and billing. PACS Sectra PACS enables radiologists to perform their work efficiently in a distributed reading scenario independently of network capacity. It is designed for variations in network quality and works efficiently over high-latency networks. A central requirement for a PACS is the ability to identify and store all images together with any relevant information, and to allow image and information access to authorized users. The images, and in some cases other radiology and patient information, are typically stored in a short-term and a long-term archive, with a database server keeping track of where the image data is stored. End users typically access the PACS through workstations, which are connected to the server components through a data network.


Figure 3: PACS building blocks Source: Sectra


Test results Once the core test environment (as described above) was built, all Sectra-supplied PACS and RIS virtual machine images were imported into the two-node ESX cluster. Each virtual machine was brought online and the applications were tested for basic functionality. This involved testing communication, database operations, importing images to the ImageServer, and ensuring that images were displayed correctly through the Sectra PACS GUI. Once all virtual machines were functioning and the core environment was fully validated, approximately 6 to 7 million images were loaded onto the LUNs that were presented from the back end of the NS-960. On completion of the data loading, each test case was then executed. A description of the test cases and the resulting observations are contained below.

VMware HA At the beginning of this test case, the virtual machines were running across two ESX servers. One ESX server had a reboot command issued to it from vCenter. All virtual machines that were hosted on that server correctly moved to the second server in the cluster and each Windows virtual machine restarted as expected. Once all services were validated as having started and the database was available, the application was fully operational again. The entire failover process took approximately five minutes from the point that the ESX was shut down.

VMware vMotion All virtual machines were running and a cine loop test was running on the ImageServer. This test involves using the application to read approximately 2,500 images from the same scan so that the images display in movie style. Before initiating VMware vMotion, the images were displaying to the screen at a rate of 15 to 25 frames per second (fps). vMotion was first run on the ImageServer/S VM. The application continued to operate throughout and was displaying images on the loop. The lowest fps that was recorded was eight for a period of approximately 20 seconds while vMotion was running. ImageServer/S migration during image import showed no apparent degradation of performance. The virtual machine relocated in less than two minutes. The test was repeated for each of the following: Sectra Healthcare database (SHDB), Workflow Information Service Engine (WISE), and Sectra Healthcare Server (SHS). The behavior observed for each virtual machine was consistent with the ImageServer – SHDB: 30-second window of eight to 10 fps; WISE: 30-second window of five to 10 fps and SHS: 15-second window of five fps. This test showed that vMotion can operate while users view images without disruption.

VMware DRS A third ESX node was added to the cluster. This Dell R710 was zoned to the native FC ports on the back end of the NS-960 and had the same storage presented to it as the other two nodes. A heavy load was generated on the ImageServer virtual machines by ingesting images from a Linux virtual machine as well as running cine loop testing from a workstation that was accessing images stored on the ImageServer virtual machines. Once a sufficiently heavy load was reached, the DRS policy correctly and automatically moved the virtual machine from one ESX in the cluster to another. The results and behavior observed here were consistent with the previous vMotion testing. Note from this test case how the intelligence of the ESX automatically executes the vMotion at the appropriate time.


LUN migration A 1 TB RAID 5 (4+1) Fibre Channel LUN was presented as an RDM to the ImageServer. The application showed 15-20 fps when the cine loop test was executing in this configuration. The workstation being used to run the Sectra PACS GUI and view the images at that rate, was reading the images from the ImageServer drive (a 1 TB FC LUN of ID 401). Simultaneously, a Linux VM client was ingesting images to the same ImageServer drive. This behavior simulated a real-life scenario of new scans being written to the system while existing cases were being reviewed by healthcare professionals. The read and write I/O throughput that was generated from this activity on the back end of the NS-960 was captured and can be seen in Figure 4 below. The peaks and troughs in the graph are as a result of the variety in the number of images associated with each case being reviewed and whether it was being viewed in 3D mode or not. Using Unisphere, a LUN migration operation was initiated to move the 1 TB LUN from FC to EFD while the cine loop test was executing. There was no negative impact to the application while the migration was in process.

Figure 4: Read I/Os per second – Fibre Channel


Once the migration to RAID 5 (4+1) EFD was complete, the application displayed images at a rate of 25-40 fps. The same testing that was performed on the FC LUN was repeated for the EFD LUN. Once again, the read and write I/O throughput was captured and can be seen in Figure 5 below. It can be observed that there was a significant increase in I/O from when the same test was run on FC drives.

Figure 5: Read I/Os per second – EFD

Celerra NAS share for archiving A CIFS server was created on the NS-960 and a single file system was presented as a CIFS share to that server. The share was mapped as a network drive on the ImageServer. The archiving feature of the Sectra application was modified to point to that share as the location to move archived files. The NAS share was able to function as an archive. The Celerra, once added as the archiving destination, introduced no performance issues or limitations on how fast the PACS application could archive the data to the share.


Celerra Replicator Celerra Replicator V2 was set up between the NS-960 and the NS-120 for the archiving file system that was used in the previous test case. While the share was being used as a destination for archived images, a switchover command was issued from Unisphere. The share that was previously write-disabled on the NS-120 instantly became available as read/write and all data was accessible. Although not configured for the purpose of this test case, the use of Virtual Data Movers to replicate the CIFS information is a feature that could be considered for installations in the field.

EMC MirrorView/S A MirrorView/S mirrored pair was configured on the 1 TB FC LUNs that were presented to the ImageServer/FS and ImageServer/S. Once the initial synchronization was complete, the primary mirror shows as active and the secondary shows as synchronized. To simulate failure, the ImageServer virtual machines were shut down, the RDMs were removed from the virtual machines, and the LUNs were removed from the storage group. Using Unisphere, the secondary mirror was promoted. This then became available to be mounted as a read/write LUN on the relevant virtual machine. The LUNs were then added to the storage group and made available to the virtual machines as RDMs. Once the test was complete, it was observed that all data was again available from the LUNs that were presented from the back end of the NS-120.

Celerra File Level Retention Celerra File Level Retention policy (FLR-Compliance) was enabled on the NAS file system at creation time. This was the file system that was used for archiving. Retention policies were set on files/folders using the FLR toolkit, which was installed on one of the virtual machine Windows servers. Distributed Hierarchical Storage Management (DHSM) was enabled and a user was created on the Celerra to allow authentication against the Celerra when connecting from the toolkit. It was not possible to then edit or delete the files until the set retention time/date had passed. Once the time passed, the file was once again available for modification. Note DHSM is a policy-based system that is used to determine where files should be physically stored. In most cases, policies are based on file size and/or last access time and are used to identify data that can be moved to slower, less-expensive storage.

iSCSI iSCSI LUNs were created on the Celerra NS-960 and presented to the ImageServer virtual machines. The ImageServers had MS iSCSI Initiator software installed and were able to connect to the Celerra Data Mover target. Image data was ingested onto the iSCSI LUNs and the Sectra PACS application was used to view images through the workstation as described earlier. It was noted that the iSCSI storage was able to provide the required functionality for the application. Although the environment hardware limited the number of ports that could be used for iSCSI, it was shown that the cine loop test was at 15-25 fps when no archiving was occurring and 15-20 when archiving was taking place.


Conclusion IT is being asked to meet the increasing needs of three very important, demanding and different constituents: clinical end users, regulatory bodies, and senior management. Clinicians need quick access to medical images that are always available. Regulatory requirements are demanding information protection and quick recovery in the event of a disaster or interruption to the lines of business and senior management is asking IT to reduce capital investments, operating costs, and the associated complexity built into their infrastructures today. As a result Sectra PACS must run in a virtualized environment and leverage the tools and technologies that can deliver the levels of performance, availability, and compliance, while at the same time be able to deliver increased productivity and cost reductions. Testing was conducted on a Sectra PACS customer environment with replicated EMC unified storage platforms (NS-960 and NS-120). Testing showed how:

• EMC tiered-networked storage platform provides high availability, reliability, scalability, and performance. Employing a small number of Flash drives on the array can provide a better performing and cost-effective alternative to employing traditional practices of high spindle counts and short-stroking disks to service I/O requirements. With the addition of EMC EFDs read throughput was nearly double than on Fibre Channel drives, while write throughput remained fairly constant on EFDs and significantly improved over FC.

• Sectra PACS runs in a VMware virtualized environment and provides high availability, disaster

recovery responsiveness. This results in IT cost savings due to the ability to treat the data center as single pool of processing. At the beginning of this test case, the virtual machines were running across two ESX servers. One ESX server had a reboot command issued to it from vCenter. All virtual machines that were hosted on that server correctly moved to the second server in the cluster and each Windows virtual machine restarted as expected. Once all services were validated as having started and the database was available, the application was fully operational again. The entire failover process took approximately five minutes from the simulated point of failure. vMotion was shown to operate while users viewed images without disruption or significant performance degradation. A MirrorView/S pair was configured on the 1 TB FC LUNs that were presented to the ImageServer/FS and ImageServer/S. Using Unisphere, the secondary mirror was promoted. This then became available to be mounted as a read/write LUN on the relevant virtual machine. The LUNs were then added to the storage group and made available to the virtual machines as RDMs. Once complete, all data was available from the LUNs that were presented from the NS-120.

• With EMC unified storage and Unisphere management capabilities, expanding pools to cover

growing capacity and I/O requirements has been simplified. Using Unisphere, a LUN migration operation was initiated to move the 1 TB LUN from FC to EFD while the cine loop test was executing. There was no negative impact to the application while the migration was in process.

• EMC information management and protection software helps attain recovery point and recovery time objectives for business continuity and disaster recovery plans.

• Active archiving components provide secure and rapid recall of historical images with Sectra PACS.


The archiving feature of the Sectra application was modified to point to that share as the location to move archived files. The NAS share was able to function as an archive. The Celerra, once added as the archiving destination, introduced no performance issues or limitations on how fast the PACS application could archive the data to the share. Celerra Replicator V2 was set up between the NS-960 and the NS-120 for the archiving file system that was used in the previous test case. While the share was being used as a destination for archived images, a switchover command was issued from Unisphere. The share that was previously write-disabled on the NS-120 instantly became available as read/write and all data was accessible. Celerra File Level Retention policy (FLR-Compliance) was enabled on the NAS file system at creation time. This was the file system that was used for archiving. Retention policies were set on files/folders using the FLR toolkit, which was installed on one of the virtual machine Windows servers.

References For additional information, see the documentation below.

• MirrorView Knowledgebook: FLARE 30 - A Detailed Review

• EMC Celerra Network Server Documentation

Virtualizing Sectra RIS/PACS Solutions for Healthcare ... · Virtualizing Sectra RIS/PACS Solutions...

Documents

Transcript of Virtualizing Sectra RIS/PACS Solutions for Healthcare ... · Virtualizing Sectra RIS/PACS Solutions...