Video surveillance, a key storage challenge for many law enforcement agencies, became the entry point into hyper-convergence...
for the Bergen County Sheriff's Office in New Jersey.
BCSO began using hyper-convergence before it was even called that. Eight years ago the agency installed Pivot3 CloudBanks storage for video surveillance systems. CloudBanks combined compute and storage into one unit before the term hyper-convergence was coined. Philip Lisk, director of IT for BCSO in Hackensack, N.J., said the agency started with 50 cameras and today has more than 3,000 in its jails and courthouse.
Five years ago BCSO moved from CloudBanks to its Pivot3 vSTAC OS hyper-converged systems during a video upgrade in its jails. That took the agency from 500 TB of storage to more than a petabyte. Now it is adding another 400 video cameras and 30 hyper-converged vSTAC Pivot3 Enterprise hyper-converged infrastructure appliance clusters that will expand its storage capacity another 1.5 PB. The new hyper-converged appliances will also handle BCSO's virtual desktop infrastructure (VDI) storage. The 3,000 cameras connected to Pivot3 vSTAC storage do not include the cameras in patrol cars. Those are stored on a separate system that Lisk said he plans to integrate onto Pivot3 storage in 2017.
BCSO stores data from its cameras for 60 to 90 days. Lisk said his office used to replicate data to protect it, but there is too much data now. It relies on Pivot3 vSTAC for disaster recovery as well as VDI and long-term storage.
"We got into hyper-convergence because of the amount of data we ingest," Lisk said. "Traditional methods of backing up to tape and other sites, and then copying, is becoming increasingly taxing. There's no way to get it done effectively. We used to be able to record data, store data and replicate data. That is no longer the case. There's no way to replicate all that data. We need data that is online all the time."
Lisk said he looked at other hyper-converged and more traditional storage systems for his current upgrade but stuck with Pivot3. He said one of Pivot3's selling points is its redundancy, which saved BCSO at least once during a disaster.
"You can actually lose a complete server without losing any data or any of the resources for the end user," he said of the vSTAC systems. "That's one of the big things for us: The data is guaranteed to be there. That's very important in video surveillance. You don't want to say, 'We don't have that data because the server was broken that day.'
"We had an incident about a year ago where we had a water leak and water damaged a server. We took the server out, put a new one in, put it back into the appliance stack and users were none the wiser. We didn't lose one bit of data."
Lisk said BCSO still has data on application servers that it will migrate to Pivot3. Eventually, the Pivot3 systems will include all storage for VDI, Microsoft Windows servers and Windows documents.
"We'll eliminate several single appliance servers and move them into our VMware environment powered by Pivot3 hardware," he said.
He said he is looking forward to using some of the new management features Pivot3 has added. "They just brought out some things we talked about with them late last year," he said. "They added a single pane of glass for management so I can see everything across all the arrays I have. They also added advanced reporting on alerts for things that are OK. I was always able to get alerts when things weren't working. But you're wondering, 'Did I miss an alert?' Now I can get the alert telling me everything is good."
Hyper-convergence newcomers search for their place in the market
Why hyper-converged architectures are still in their infancy
Software-only hyper-convergence requires data center knowledge