Safeguarding data in the enterprise

Server virtualisation and managing growth emerged as common themes when Ulrika Hedquist spoke to local users about storage and disaster recovery

Webdrive

For Auckland-based internet services and hosting company Webdrive, finding a storage solution to keep pace with the growth of the business was a challenge. The company has grown on average 65 percent each year for the past three years.

Webdrive, which looks after around 80,000 domain names and tens of thousands of websites, was running out of capacity, says the business development manager, Robin Dickie.

“We simply outgrew our existing SAN [storage area network] storage,” he says. “And SANs aren’t exactly cheap, so it was quite a problem for us.”

There were also backup issues before the upgrade, he says. Webdrive is expected to backup customers’ data every day and have it available to restore. However, in some cases backups were taking more than a day to run, he says.

“If you are running a daily backup and it takes more than 24 hours, you have got a major issue.”

Webdrive had a look around the market and decided to move to an Oracle unified storage system, consisting of two Sun 7420 boxes and a Sun 7120. The main reason for going Oracle, was the ability it offered to scale to a larger size and in a modular way, says Dickie. “Webdrive has terabytes (TB) of data, but it’s not hundreds of TB, however it is heavily accessed – every day.

“We have found we don’t tend to run out of storage space, we tend to run out of the ability to access that storage.”

The unified storage system has worked well for the company, says Dickie. If more capacity is needed, it is easy to add modular storage to increase performance. Among the benefits is that staff can manage and look into performance issues, he says.

“We can drill down to a specific file on a website that is causing a problem, basically in real-time, which is hugely beneficial to us,” he says.

While Webdrive is using about 5TB of data at present, the platform can go up to one petabyte of storage.

“This should meet our needs for the foreseeable future,” he says. “This is a big investment so for us, so it is important that it is not something needing to be replaced in two years.”

The unified storage platform also offered benefits in the disaster recovery area. Previously, the company had a single SAN. If that had failed it would have taken about three weeks to move the data onto another SAN.

“That would have been an absolute disaster,” says Dickie. “That would have been three weeks without people being able to access websites.”

Now, the company is using two Sun 7420 servers and a 7120, meaning if one of the boxes fails, the other one will pick up the load and, “we wouldn’t miss a beat”, he says. Backing up data is now instantaneous, compared to hours or even days previously.

According to Dickie, when it comes to storage, the biggest challenge is cost.

“It is a big investment and there is risk involved. Buying the wrong technology means wasting potentially hundreds of thousands of dollars.”

Capacity isn’t the only consideration; the ability to use that capacity is also vital, as well as the ability to manage, move and backup data, he says.

Ministry of Agriculture and Forestry

Facing an inevitable hardware upgrade, the Ministry of Agriculture and Forestry considered its options.

“We wanted to explore new technology and cloud enablement, as government agencies look to increase collaboration,” says MAF’s deputy director of general business services, Nigel Prince.

A year ago MAF decided to swap its hardware for outsourced infrastructure services. The government department signed up with infrastructure provider Revera to virtualise 200 servers and SANs.

As part of this transformation, MAF shifted to a storage-as-a-service model, says Bryce Johnson, MAF’s core services manager. Rather than leasing and owning its own SANs, MAF is now buying storage on a “per terabyte per month” basis, he says.

“Previously, we had a lot of money tied up in kit that we didn’t need the full capacity of,” says Johnson. “Storage-as-a-service means all we do is pay for the storage we are actually using. We can increase it and decrease it [as required].”

The solution is beneficial for transparency of costs, he says.

“If the business adds an application and it needs 500GB of storage, we can tell them exactly what that is going to cost.”

And if the business decides it doesn’t need the app anymore, the minute it is turned off so are costs, he says.

“There is no waste,” says Prince. “Internally we can transfer costs to the people who actually use the resources. Cost allocation becomes much clearer, and we get a better coupling of business decision and cost and consequences.”

The shift to infrastructure-as-a-service will help MAF avoid hardware capital, lease and management cost. This is expected to shave millions of dollars from MAF’s IT operations budget, says Prince.

“We also don’t have to worry about upgrading the SAN or buying new ones when they wear out, or keeping the software up-to-date,” says Johnson. “That’s all Revera’s problem.”

As part of the service, Revera also provides the equivalent space to hold backups for MAF, meaning the organisation can restore rapidly, he says.

If you decide to go for the outsourced model, you have got to know exactly who you are dealing with, he warns.

“You need to make sure your provider has got the technical expertise to do it,” says Johnson. “It comes down to administration. They need to be able to accurately attribute the storage against all your servers and then invoice you appropriately.”

St John

Medical and emergency ambulance services provider St John has gone for a Hewlett-Packard EVA (Enterprise Virtual Array) storage solution. St John has been using HP for the past five years, says Ray Pitch, acting ICT operations manager at St John.

The not-for-profit and membership-based organisation has two datacentres, one in Auckland and one in Christchurch, that effectively mirror each other, says Pitch. Together, the datacentres hold around 50 TB of data.

Over the past two years St John has consolidated from five regions to one shared services organisation, where the national office in Wellington now runs the entire ICT operation. Bringing all the systems together required a flexible storage system, says Pitch. St John also wanted a system where it could move processing within a site, as well as between sites, should it need to.

“That strategy led us down the road of virtualisation, which requires phenomenal amounts of storage,” he says.

The HP StorageWorks solution offered the reliability and flexibility benefits St John was after, as well as cost-benefits, he says. Naturally, cost was one of the main objectives for the charitable organisation.

For the commercial services such as payroll, HR and finance systems, St John is achieving 99.97 percent reliability, he says. As well, the ICT team is able to move systems around if they start to detect equipment failure.

“Hardware failure is now a thing of the past. It is not a cause for our systems to go down anymore,” says Pitch.

With St John moving towards a stronger focus on preventative health, the organisation has a number of new projects on the go, he says.

“We are now able to provision servers for projects within hours. We have sufficient server capacity, so we don’t have to go out and buy a brand new box.”

There is also additional server capacity for testing, he says. As an example, St John recently replicated all its financial systems and ambulance transport system in order to test the new GST increases, he says.

It took St John longer than planned to implement the storage system, mainly due to lack of detailed documentation as to how the existing systems were interrelated, says Pitch. Among the lessons learnt is to allow enough time for auditing what you have currently got, he says.

“There are also a lot of things to be learnt about performance tuning, especially when are you geographically spread. Version-control among all the components becomes quite critical for performance.”

Future storage plans for St John involve continued expansion.

“Our challenge now is physical space for additional racks,” Pitch says.

Kiwibank

Kiwibank decided to move its mission-critical production systems from Wellington to Auckland and is currently in the first phase of the process. “To do that we needed to embark on a virtualisation strategy,” says Kiwibank general manager of IT and business delivery, Ron van de Riet.

The drivers behind the strategy were to make management of the shift easier and to better manage growth, he says. Being a fast growing organisation creates a high-change environment. The bank’s rapid growth has meant it has built infrastructure solutions on the go, which led to a plethora of SANs and equipment, he says.

Kiwibank started a search for a storage and virtualisation solution, and as part of that visited sites that were using enterprise-level SANs.

“We looked at some big players, including high-risk organisations in Australia before we made our decision,” he says.

A Hitachi system consisting of two Hitachi Universal Storage Platform VMs and Adaptable Modular Storage 2500 was chosen, which has been running for the past six months.

“Hitachi stood out because they could virtualise all our existing SANs under one virtualisation model, which makes management so much easier,” says van de Riet.

Resilience was also important to the bank.

“We needed absolute stability,” he says. “We didn’t want to be ‘bleeding edge’. We wanted something well-proven that could scale as the bank grows.”

Through virtualisation, the Hitachi system is now enabling Kiwibank to start migrating SANs without needing to take systems down, and without impacting customers, says van de Riet.

Another reason Kiwibank selected Hitachi is the company has a strong engineering culture, he says.

“The manufacturing capability is just amazing. It’s quite humbling when you see the enormity of it. It gives you a lot of confidence that the company is going to absolutely deliver what you are looking for.”

Christchurch Polytechnic Institute of Technology

Christchurch Polytechnic Institute of Technology needed a new storage solution in 2008. Its existing SAN was reaching end of life and the vendor was ceasing support for it. With about 25,000 tertiary students per year and 2000 staff, CPIT is the fourth-largest polytechnic in New Zealand and the largest in the South Island.

CPIT sought a new solution to support a heterogeneous IT environment with Microsoft, VMware, Apple, Linux and Novell in the mix. Capacity, robustness, expandability, future procurement terms and availability of support in Christchurch were also among the requirements. The ability to scale was particularly important.

“It is quite feasible that one of our academic staff could receive a significant grant and say to us ‘I need four terabytes of storage quickly’,” says the polytechnic’s ICT division director, Mark Marshall.

“We needed a system that was agile and could expand well beyond our predicted needs. This was to ensure we did not have to purchase another discrete storage system and increase our costs and management load as a result.”

The polytech chose a EMC CLARiiON CX4-240 midrange networked storage system, offering up to 231TB of capacity and 99.99 percent availability, which met CPIT’s uptime, performance and scalability requirements.

A strong benefit of the solution is that managers now have access to real-time financial analysis and reporting information, stored in a data warehouse on the SAN, rather than having to review the last PDF generated from the CPIT accounting system, says Marshall.

The system is also helping the polytech’s IT department to respond to demand from the institution quicker, essentially because the new system is three times the size of the previous SAN. The system also frees up time for Marshall and his team to focus on other tasks, as it requires minimal hands-on management, he says.

Join the newsletter!

Error: Please check your email address.

Tags mafdisaster recoverystoragewebdriveSpecial IDst john

Show Comments

Market Place

[]