Bellway Homes was founded in 1946 in Newcastle upon Tyne as a family-run housebuilding firm, and since then has grown to one of the most successful residential property developers in the UK. They have a strong focus on quality, trust and service and now have over nineteen regional divisions hiring local talent.
For Bellway, emails are critical to their day-to-day work processes. If the system went down for five minutes, it would impact their whole business. There are a lot of time critical communications for land and property purchases, for getting bids in at the right time, and for ordering vital parts when they are needed.
Martin Reah, Senior Infrastructure Engineer of Perfect Image, says, “It is imperative that Bellway users are able to send emails at all times. Delays in sending emails can have a negative knock-on effect to orders, sales and land purchases.”
The company was previously experiencing problems as email servers were running on a virtualised platform with shared storage. Space was rapidly running out because the constant increase in emails storage had used up all of the available spare disk space on the existing storage area network. The company was considering buying more storage in and extending its on-site server’s storage.
Bellway used to pay upfront for the servers and the maintenance and support. To extend, they would have had to pay for the kit, to house it in someone else’s datacentre, and on top of that pay upfront costs of around £50,000 to extend the hardware.
They needed a way of reducing their costs in the short and long term, in order to invest in the more valuable areas of the business.
Having a failover process that was complicated and took a long time to complete wasn’t going to be enough if the worst were to happen. Although resilience was built in, because of the version of Microsoft Exchange they were running, it didn’t always work as it should.
The company wanted a way of reliably accessing their emails with high availability, business continuity and improved backups.
Weekly onsite backups previously took two and a half days to complete but were unreliable as to whether they would work the first time or whether they would need another two and a half day attempt at a backup. These backups were stored on multiple physical tapes that were costly, difficult to manage, and time-consuming to arrange. This was using up valuable operational time.
Bellway needed a consistent process that would work every time and would not require physical tapes to store their backup.
In January 2017 Bellway went out to tender to see what IT providers thought they should do with their emails. Many companies replied with solutions that were paying to extend the storage of their current servers and remaining onsite.
Perfect Image came forward with a solution utilising the cloud’s flexibility, unlimited storage, and monthly payments. Microsoft Office 365 had been considered but because Bellway has over 2,500 mailboxes they were at the point where it would be more cost-effective to host emails themselves.
We thought about the short and long-term implications of the solution, as we didn’t want to continue adding extra disc space at a high cost to the company – instead, we decided on something that was more flexible, scalable and resilient and would better meet their business needs.
Martin explains, “We weren’t really given a specification, Bellway just wanted emails to work better and be more reliable. We chose AWS because at the time the hosting solution on Azure was nowhere near comparable to AWS. Also a lot of Bellway’s existing infrastructure was already hosted by Amazon, so it just made sense.”
The work had to be undertaken swiftly because Bellway was running out of space to store and send emails. Perfect Image quickly migrated a portion of the users into the new solution to ease the strain, and the urgency immediately dropped.
To begin, Perfect Image upgraded the company’s Microsoft Exchange from 2010 to the 2013 version to work better with load balancing requirements (to distribute data across networks, applications and services to enhance reliability and performance) and increase backup options. We then built the new Exchange server using Microsoft best practices into Amazon.
The three main elements of building the solution were to use load balancers (for improved scalability and robust security), a Virtual Tape Library or VTL (for cheaper and more efficient backups), and Virtual Machines split across multiple geographical locations for greater resilience for local communications and power failures (reducing the need for expensive physical hardware systems). Then we could migrate their emails and relevant services across.
The system is now highly reliable, accessible, and available. Perfect Image complete regular testing to ensure the disaster recovery is as reliable as it needed to be.
Previously, disaster recovery to get the system back up and running would have taken around 8 hours, and still had the possibility it might not work. Now, that process takes just fifteen minutes.
Sean Lally, Director of Information Technology at Bellway, says, “Our email system can now be recovered in just fifteen minutes and that gives us real peace of mind, because they are so vital to our business. In the small chance our system fails, we know we’ll be back up and running in no time.”
Costs have gone down and are more convenient since the new system was implemented for many reasons:
Sean says, “Costs are broken up into manageable monthly payments. This really helped us to get started with the solution in the first place, knowing there would be no huge upfront expenses for the cloud.”
Instead of unreliable physical backups, the company now used a Virtual Tape Library. Now, they don’t have to change tapes, more of the process is automated and therefore required less manual effort, and was less complicated and time-consuming, and it was web-based to remove the need for physical tapes. Due to the way it has been designed, it doesn’t matter how long backups take to run as they don’t fail – there is a much higher success rate to almost 100%.
The company is no longer restricted in any way by their email system.
If they want to add a further 10,000 employees to their email system, they can, by easily scaling it to add in as many mailbox servers and components as they want.
If their emails grow too large and users don’t want to delete emails, we can create an archive and move emails with custom policies. This will double the amount of email storage per user.
Sean explains, “Whatever the company needs in the future, Amazon can handle it. The storage on AWS is infinite, so we don’t have to buy more hardware. When we need more space, Perfect Image arranges for more from Amazon, which has already happened as we continue to grow.”
As Amazon Partners, we decided to put the workload into Amazon and part of the design was to make it highly available. This technology was critical, but to implement that on their previous on-site system it would have cost upwards of £30,000. In comparison, Amazon’s equivalent was extremely cost-effective, as it only required the running costs of a virtual machine.
Martin says, “We’ve got a few load balancers making sure that if one side of the system goes down, or if any component fails, then it should be able to repair itself. The end user shouldn’t even know it’s gone. That’s one of the reasons we used AWS, because at the time Azure didn’t have any kind of equivalent load balancer technology.”
Using Amazon’s load balancers, a device between their Exchange servers was utilised so that if one server went down, it would switch to the other that was live. Even if one part of the infrastructure failed, their workload wouldn’t be affected – there wouldn’t be a complete loss of services as there would have previously.
Sean concludes, “The main benefit of the new system is its reliability. Our on-site email system could be somewhat unpredictable, but with our upgraded AWS system, no one worries about it. We just know it’s going to work.”
The Amazon Web Services infrastructure in place will soon be expanding. Perfect Image has started the process of implementing highly available SQL onto the system to run databases and production workloads. The SQL server is currently in development and being used on a trial basis, with a view to use it with their new ERP system for a reports server.