by July 4, 2012 0 comments

Infosys goes for a DR & BCP solution to make sure their business won’t be affected even in case of a data center disaster and systems would be up and running in minimal time without manual intervention

Infosys provides business consulting, technology, engineering and outsourcing services to clients in over 30 countries. They have over 900 physical and virtual servers with more than 30 million files and heterogeneous technology platforms like Microsoft SQL Server, .Net, SAP, Weblogic, Skelta, Symantec and many more that comprise the digital business engine of Infosys.

[image_library_tag 509/70509, border=”0″ align=”middle” hspace=”4″ vspace=”4″ ,default][image_library_tag 510/70510, border=”0″ align=”right” hspace=”4″ vspace=”4″ ,default]

All these files were stored in a 2-data center design cluster that consisted of storage sub-systems in both data centers. Nodes of clusters were distributed across two data centers and shared disks mapped to a cluster were FC-connected to all the nodes across data centers. In order to tackle disaster recovery for storage subsystem disaster, disks from one data center were synchronously copied to the second data center.

Reliance on manual intervention to initiate recovery during disaster was prone to errors. Also, it was slow and time taking with RTO coming at around 32 hours. This prolonged backup window was causing performance issues during business hours and fear of potential data loss, reliability, data protection & scalability issues for file shares was also disturbing.

The implementation

The process of deployment began with Hitachi Storage Cluster (HSC) deployment, which intends to break the conventional shared storage model that is widely followed across the globe. It was followed by Hitachi Storage Cluster (HSC) deployment to synchronously replicate data to the second data center automatically with an RTO of less than an hour compared to the previous RTO of 32 hours. This has ensured that systems would be up and running in minimal time without manual intervention.

Then Metro Cluster over HNAS was deployed for file shares. It comprises of two active clustered file share setups with storage level synchronous replication from primary storage in data center A to secondary storage in data center B, and has solved the site level disaster recovery problem for file shares along with near zero RPO now, as compared to RPO of over an hour. The last deployment is of Hitachi Content Platform on top of HNAS metro cluster, which is a feature rich, multipurpose content storage solution meant to handle the complexity of mixed storage environments.

What other technologies were considered and rejected during the deployment?
[image_library_tag 511/70511, border="0" align="right" hspace="4" vspace="4" ,default]

The need of the the hour was to “Do More with Less”. We leveraged our existing investment on highly virtualized storage and server private cloud which had already proven itself on performance and scalability fronts by providing 100% uptime for many years. The selected solution protected our existing investment, met the business requirements and provides additional value added features.

What is unique about this deployment?
We have conceptualized, designed and deployed a DR Solution that automates disaster recovery for failures up to a data center level failure. The solution was delivered with minimal Capex, thus protecting the existing investment. Other key USPs include: First of its kind large scale Hitachi Storage Cluster deployment for data base clusters in the world; first ever Hitachi Metro Cluster implementation in production environment in the world; first Hitachi Content Platform deployment with 3-site architecture in APAC.

[image_library_tag 512/70512, border="0" align="right" hspace="4" vspace="4" ,default]

An intellectual property?

While designing DR solution, Infosys came up with a unique idea of providng all current HSC benefits with centralized configuration management. This will have custom developed menu for controlled failover and failback mechanism. This solution can bring value addition to the rest of the customers who want to deploy DR solution using Hitachi Storage. The second part extends the scope of HSC to virtual environments. Existing version of VMware has some limitations with MS cluster configured in virtual environment. This solution helps overcome this limitation and enhance DR capability, so they decided to file a patent.

The Result

[image_library_tag 513/70513, border="0" align="left" hspace="4" vspace="4" ,default]

The deployment has resulted in significant reduction in RTO from 32 hours to less than an hour for the entire landscape in case of a site level disaster. Manual level intervention for initiating the recovery process in case of a site level disaster has been replaced by a much needed fool-proof DR solution for mission-critical applications. The RPO has been reduced from 24 hours to near zero for file share services along with reduction in backup window and savings on storage costs by 50%+ and 75%+ respectively. The deployment of Hitachi Content Platform on top of HNAS metro cluster offers a feature rich, multipurpose content storage solution with security, scale and broad application support. It improves the dispensation of a wide range of IT services, simplifies storage, tracking, and access to unstructured data.

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.

Your data will be safe!Your e-mail address will not be published. Also other data will not be shared with third person.