Case Study: HarperCollins
- Legacy SAN not meeting business requirements
- Adhering to parent company’s ‘Cloud First’ policy
- Migrating an enterprise infrastructure into the cloud
Commvault data protection and disaster recovery inside Microsoft Azure cloud and on-premises.
- Modernization and consolidation of IT by leveraging Microsoft Azure
- Move to a more manageable OPEX model
- Ability to leverage existing Commvault investment for cloud migration, and ongoing protection on-premises and inside Azure
- Multiple restore options and secure public cloud disaster recovery
- Deduplication delivering high savings in both space and IT department’s time
With a heritage stretching back nearly 200 years, HarperCollins Publishers LLC is one of the world’s largest publishing companies and part of the “Big Five” English-language publishing companies. Founded in Glasgow in 1819, William Collins & Sons merged with New York’s Harper & Row in 1990 to create the HarperCollins worldwide brand, with News Corp as its parent company. Today, HarperCollins UK employs some 900 people in London, Glasgow and Honley and has a catalogue ranging from cutting-edge contemporary fiction to award-winning apps and everything in between. HarperCollins UK prides itself on leading the industry in innovation. It was the first trade publisher to digitise its content and create a global digital warehouse. Parent company NewsCorp is equally ahead of the curve when it comes to technology and has a “cloud first” strategy that aspires to migrating 75% of its global IT infrastructure into the cloud.
Original Commvault deployment
HarperCollins first deployed Commvault in 2011 to make the back up of 150 servers at its Hammersmith offices and 250 in Glasgow more efficient. The publishing house deployed on-premises Commvault solutions for backup, deduplication, and disk to disk (as well as disk to tape) replication for DR purposes. The legacy HP SAN system had not been meeting the company’s data requirements but Commvault reduced traffic and increased the speeds of backup and recovery through deduplication.
Eighteen months ago, NewsCorp decided to bring its subsidiary companies, Dow Jones, News UK and HarperCollins, into the same building in London. Unlike the Hammersmith premises, the London office had no data centre or back office, so there was no space for servers, (assuming the equipment could survive being moved, which was unlikely). With the company having to lease data centres, whether in the cloud or physically located elsewhere, it seemed the perfect time to implement NewsCorp’s strategy of moving 75% of IT infrastructure to the cloud.
Azure the obvious choice
HarperCollins had a very good relationship with Microsoft, so Microsoft’s Azure Platform was the obvious choice when it came to moving the data centre to the cloud and realising its cloud strategy. Microsoft brought along its partner Dotnet solutions to help HarperCollins make the transition.
In addition to following the NewsCorp Cloud First Strategy, moving to Azure cloud services enabled HarperCollins to consolidate its IT and switch to a more manageable OPEX commitment. A CAPEX model would have meant a sizeable expenditure in one go, which wouldn’t allow the business to budget as easily as Azure’s pay-as-you-go plan. Reducing the physical data centres with a cloud strategy would also save the company money on suppliers, power, and air conditioning, in addition to the physical space that racks and servers require.
“When we first moved to Azure, it was bleeding edge. There was no back up. But we had Commvault as our on-premises backup and knew that not only could it meet our requirements, but also it would integrate perfectly with Azure. So that decision was a ‘no brainer’,” said Russell Bradford, then infrastructure manager for HarperCollins, now cloud platform manager for News UK. “We were moving an enterprise infrastructure into the cloud, so we needed enterprise tools….from our experience of Commvault, we knew it was the perfect fit.”
All change – except Commvault
During the migration process, HarperCollins had to take IT personnel off their day jobs to lead the process, as well as tapping into the experience of the IT team at NewsCorp. At the time that HarperCollins was migrating to Azure, the cloud was pioneering technology. None of the IT departments were familiar with the Azure functionality or the process. It involved a steep learning curve for everyone involved, as everything would have to change – except for Commvault. HarperCollins realised the Commvault platform it had bought five years previously would work the same in the cloud as it had on-premises.
“There were worries about moving the data centre to the cloud, it was totally new technology for all of us and that was the biggest challenge,” said Bradford. “We had to test and get used to the new environment, at the same time as we were designing it. Fortunately, all the Commvault related process for back up and DR stayed the same. We took the next generation of the Commvault software, but that was the only change we made. It meant that at this highly challenging time, the back up and DR were the least disruptive.”
Aside from the ease of transition, Commvault’s deduplication feature proved to be one of the key benefits for HarperCollins, delivering high savings in terms of space and the IT department’s time.
“We have different dedupe engines, however the best one in the cloud is currently a 77.38% dedupe saving. ” said Bradford, “With previous products we had to have a dedicated resource to manage it, it would be someone’s full time job, but with Commvault, to a certain point, you could turn it on and leave it.”
As well as a small data centre leased in Telecity for log in purposes and an on-premises DC in Glasgow, HarperCollins houses its main cloud data centre in Azureâ€™s North Europe Data Centre, which is in Ireland.
It uses the Azure data centre in West Europe for DR. HarperCollins was the first European customer to get a private cloud link, called ExpressRoute, which connects it, securely and privately, to the data centre. The ExpressRoute is also far faster than a VPN.
“There were worries about cloud security, but the Azure data centre has far more security and compliance than we require. We have chosen the North European Data Centre for sovereignty reasons but, unlike other sectors, we don’t have any highly sensitive data, such as that which requires PCI compliance.”
Additional piece of mind
HarperCollins use Microsoft ExpressRoute for the critical network dependency between Telecity and Azure. This production data is protected with a dual site data protection strategy. Commvault software makes a copy directly to the primary Azure Cloud data centre in North Europe (Ireland) and an auxiliary copy to the secondary Azure Cloud data centre in West Europe (Amsterdam).
This gives HarperCollins multiple restore options in the event of any infrastructure issues or primary data loss. The design also allows for advanced disaster recovery capability in the Cloud, giving HarperCollins additional control over the protection of its data. The Azure data centre in West Europe acts as a DR site for the compute workload with server images ready to bring on-line and restore, should they need to do so, using Microsoft’s Geo-redundant storage (GRS).
“We set up the DR environment this way so that if our environment offshore went offline, we could bring it back up in Amsterdam. Because Commvault software is always copying the data across, we have complete control and complete peace of mind.”