Summary
A major medical institution with revenues ranging in billions of dollars has several high end facilities across the south of the country.
The hospital employs the services of thousands of practitioners across almost all major specializations along with tens of thousands staff members.
Strict adherence to regulatory norms under HIPAA and Sarbanes-Oxley (SOX) implied a mandatory, remotely located disaster recovery site that was capable of replication capabilities in real time
Information Systems Background
The main office and Disaster Recovery site are both located along the eastern zone, separated by approximately 1,500 km.
Daily replication results in more than 10 GB of data transmission volumes
The throughput on OC-3 WAN lines connecting to the DR site was constrained due to heavy data traffic and latency issues
Backing up all the data meant using numerous applications
WAN Optimization Solution
Consolidated throughput for data replication was nearly tripled
The performance of data replication applications was also enhanced with 80% improvements
Consistent delivery of the 5 minute RPO objective was achieved for high priority data
Bandwidth upgrade was deemed unnecessary, which saved the company more than half a million dollars annually
Data storage and protection is a mission critical objective in the healthcare sector. Stringent regulatory norms make this mandatory. Failing to comply can lead to a breach of privacy or in extreme cases, even fatalities.
The IT department at a well known medical institution in the healthcare sector had to revamp their disaster recovery solution in order to comply with HIPAA and Sarbanes-Oxley norms.
By using WAN Optimization solutions, the company was able to enhance the throughput while replicating data to the DR site without making major changes to the existing network and storage infrastructure. RPO objectives were achieved. Connectivity expenses were brought down by more than half a million dollars. Now, the organization has a resilient disaster recovery solution that is scalable and can be adapted to changing business needs.
Gaps in the Existing DR Solution
The medical institution’s IT systems managed a variety of data types. Each came with its own specific demands. Data replication was achieved through more than 50 odd applications. File, home directories and system images were stored and copied on to disks. Application based replication was also facilitated for Exchange, Active Directory and DNS.
Recovery point objectives (RPO) and recovery time objectives (RTO) varied from application to application. For instance, high priority data had RPOs and RTOs of less than five minutes. This depended largely on the disaster recovery (DR) solution’s ability to deploy system images on consignment hosts.
Systems with application based replication capabilities were deployed at the DR location and could achieve almost 0 RTO and RPO objectives. The RTOs and RPOs were a lot more lenient for less critical applications.
All applications failed to achieve their RPO objectives when data was replicated over OC-12 lines that connected the organization’s primary data center with the disaster recover location. This was partly due to extremely high data volumes, which is normally the case in DR and data replication exercises. The medical institution experienced daily transfer rates in excess of 10 Gbps.
The organization also had to deal with the problem of insufficient bandwidth availability for data transmission. Disaster recovery and data replication exercises are often constrained by connectivity issues such as latency, bandwidth and data loss that collectively impact network performance. Even in the case of OC-12 lines, WAN speeds couldn’t go past 25 Mbps per application because the two locations were separated by more than 1,500 km. And although packet loss was kept well under control, the network performance would drastically fall as low as 2 Mbps. If the organization bought an additional OC-12 connection, it would imply a monthly subscription expense of nearly $40,000. This might address the issue of insufficient bandwidth. However, the lack of WAN throughput owing to latency and packet loss would still remain.
WAN Optimization
A wide range of real time network optimization methodologies were deployed which boosted the output for all replication activities that were carried out across the WAN.
Network Acceleration
The distinguishing feature of any WAN Optimization solution is its ability to effortlessly overcome the TCP protocol’s limitations while transmitting large amounts of data. This can be achieved regardless of the distance between the main location and the DR site. Network acceleration methods such as window scaling and selective acknowledgements can make data transfers faster and more efficient. Delays in transmission due to chatty protocols such as CIFS which impacted some applications that were using the OC-12 connection were also resolved. Latency between the main location and the DR site was also kept to a minimum.
Network Integrity
Organizations can cut back on DR expenses by using more economical connections such as Internet or MPLS, although packet loss is more prevalent on these lines. This is because packet loss can be corrected in real time through forward error correction.
Furthermore, out of order packet delivery can be achieved through packet order correction (POC), which reduces and in many cases even eradicates the need for retransmission. This in turn increases application throughput.
State of the art quality of service (QoS) techniques can sort data transmission based on mission criticality and also ensure resource availability.
Network Memory
Organizations can maximize the output on their DR connections through WAN de-duplication techniques. The amount of data to be replicated can be brought down considerably through data de-duplication. The WAN Optimization device that has been deployed at each location monitors all the data that is transmitted and received so that redundant information can be made available locally. This enhances application delivery as well as WAN utilization.
Solution
Four WAN Optimization devices were deployed – two each at the main office and the DR site – in an active-active setup for continuous operability. At the main office, the WAN Optimization devices were delivered through an inline deployment within the network. An out of path deployment was followed at the DR site and the data was channeled to the WAN Optimization device through policy based routing.
Application performance across the OC-12 WAN lines improved drastically. Throughput rose by more than 80%, even during peak hours. The collective throughput for all the applications across the WAN increased by more than twice.
The organization could achieve their RPO and RTO targets without investing in costly circuits or implementing cumbersome configuration changes.
RTO and RPO targets could be met without making major changes to the enterprise networking and storage infrastructure. Migration for future DR plans is now also feasible. Operational expansion can be further supported through feasible strategies for more storage capacity and WAN Optimization without compromising on high availability.
A disaster recovery location is a crucial component of any business continuity solution. However, IT teams are often constrained by limitations in WAN capabilities. Even in the case of fast speed WAN connections, latency, packet loss and inefficient bandwidth utilization can seriously hamper crucial replication tasks and activities.
The resulting solution was high on performance. But more importantly, replicating data to an offsite location was simplified and made affordable, thanks to WAN Optimization.
See for yourself how the application works
Witness our cloud based platform’s security capabilities in action
Play around with the software and explore its features
Compare and choose a solution that’s relevant to your organization
Consult our experts and decide on a pricing mechanism
Disasters
