Blazing The Trail For VTL In The Cloud
March 29, 2023 Timothy Prickett Morgan
Making the decision to move IBM i applications to the cloud is a relatively easy endeavor. But covering all of the bases, particularly when it comes to disaster recovery and business continuity, takes some thought because even though the cloud is an ideal place to do HA and DR, not every cloud does it right. And doing it right often means using a mix of technologies and service providers.
IT managers are thinking of moving their IBM i workloads to the cloud for good reasons. They do the math, they weigh the risks and rewards. They take into account all of the fully burdened costs in their on-premises datacenter. They take a critical look at when their key programming and systems administration people might retire and make an honest assessment of how difficult it will be to replace them. And then they do the calculations to simultaneously mitigate their risks, minimize their costs as much as possible, and maximize the usefulness of the applications they both maintain and modernize to keep their companies humming.
In many cases, the math says move to the cloud or at the very least start engaging with managed services providers to babysit the Power Systems platform and the IBM i operating system on site. Sometimes the move to the cloud comes first, and the managed service providers fill in gaps, patching, securing, and monitoring systems running in the cloud. Sometimes engagements with MSPs for such services leads to a consideration of cloud infrastructure.
The former was the case at the European operations of a major multinational insurance and financial services firm, originally founded in the United States but now headquartered in the United Kingdom, the name of which cannot be divulged but which we can talk about for the purposes of illustration.
In any event, the US operations of this insurance giant, which manage the data processing needs of three different divisions of the company, have been fully managed by Ricoh USA for the past four years. Ricoh USA got much of its IBM i expertise from its 2014 acquisition of mindSHIFT Technologies, which is now part of the Ricoh IT Services division within Ricoh USA. (The parent Ricoh Company Ltd, headquartered in Japan, is famous for being a maker of copiers, printers, and cameras, but it is also a big reseller of IBM Power Systems in Japan and a user of the machinery itself.) Those IT facilities of this multinational insurance company in the United States are run from two datacenters, one in New York and the other in Minnesota, with Ricoh USA managing the systems.
The insurance company was happy with the managed services arrangement with Ricoh USA and started looking at how to better manage its operations in Europe and Asia. Europe came first, and the decision was made to close down the company datacenters in France and Germany and move their IBM i applications to Big Blue’s Power Virtual Server cloud infrastructure located in its datacenters in London, England and Frankfurt, Germany. These facilities have a total of 14 logical partitions running in the European datacenters, which run the core insurance applications as well as other back-office systems such as financial and ERP, that needed to be rehosted.
The migration was done in conjunction with Kyndryl, the services company spun out of IBM Global Services, which in turn brought in Ricoh IT Services to do the managed services to take care of those partitions running on IBM’s PowerVS, including backup and recovery.
Insurance companies and other financial institutions have a legal requirement to keep seven years of data online and backed up so it can be audited at any time. As everyone is well aware at this point, IBM’s Backup Recovery Media Services (BRMS), while a fine program for modest use, is not particularly good as a backup tool for systems with more than 2 TB of capacity. It just stops working at an acceptable speed. Moreover, BRMS does not have de-duplication, which is important to speed up backups across the network, as well as decrease the amount of backup storage needed. The European arm of this multinational insurance company was starting at 30 TB and looking to expand up to 50 TB and beyond, so clearly backup performance and storage costs would be important.
Thanks to the partnership between IBM and FalconStor Software that was announced last August, FalconStor’s StorSafe VTL technology is the default backup and recovery tool for IBM i instances running on the Power VS cloud service, which can handle such big backups and do recoveries to the cloud instances with ease. That partnership was demonstrated in some collaboration work done by IBM and FalconStor on behalf of Tupperware, another IBM i shop, that wanted to move its European operations to the IBM Cloud in Frankfurt and that also knew it needed VTL in the cloud because BRMS just can’t handle the load.
Now that the European operations of the insurance company are up and running, with full backups and recovery from them in the PowerVS cloud, Ricoh IT Services has a follow on project through Kyndryl to convert the seven years of old tapes into the StorSafe VTL format so it can all be accessible online and from IBM’s Cloud Object Storage without any physical tapes. Incidentally, Tupperware told us the same thing: StorSafe VTL is an excellent tool for migrating all those aging physical tapes taking up space and deteriorating, to the cloud for easier and better management.
It turns out this work on behalf of this European part of the multinational insurance company is only the beginning for Ricoh IT Services with FalconStor StorSafe VTL. They estimate they have a couple hundred manage services customers today using Dell/EMC’s Data Domain deduplicating storage systems that are excellent targets to modernize with migration to the cloud and backup using StorSafe VTL.
“It’s a practice we anticipate offering widely,” says Nicholas Mattera, vice president of hosting at Ricoh IT Services. “When we have an opportunity for a hosted Power Systems environment or a managed services Power cloud environment, the IBM Cloud with StorSafe VTL is a good option. That’s the direction we intend to take.”
This content is sponsored by FalconStor Software.
Like Tupperware, Think About Storage When You Move To The Cloud
Big Blue Taps FalconStor To Move Data To PowerVS Cloud And Protect It
Kill Three Cloud Scenario Birds With One StorSafe Stone
Former Power Systems GM Joins FalconStor For IBM i Push
Ricoh Gets Into MSP Business With mindSHIFT Buy
Best Buy to Acquire IBM i Cloud Provider mindSHIFT
At mindSHIFT, IBM i Hosting Options Abound
Tim, I do not think you understand how BRMS works or for that matter how IB i backups work. BRMS is a backup application that uses different methods to do backups. It saves data to a VTL or to physical tape or to the cloud. The Falconstor VTL does not replace BRMS, but rather is a target for BRMS to send saved data too. The Falconstor VTL has no option to run or schedule backups on the IBMi . Users will still use BRMS to save their data to the VTL.
As a side note, there is no restriction on saving more than 2 TB using BRMs and ICC to the cloud. We have only stated this as a recommendation due to possible space and transfer times requirements. If a user wants to save 10 TB to the cloud and have the space and time, they can. Nothing will stop them from doing this.
Thanks for the ping. My understanding–and this came from the PowerVS people at IBM–is that performance about 2 TB is precisely the issue. And that is why everyone is looking for something else. They may use BRMS as a front end to VTL, as I think you are suggesting, but BRMS itself is too slow on large datasets.
These are not my claims, but those of the people I have interviewed at service providers and at IBM.
Retired IBMer here, still enjoy keeping up on IBM i stuff. I was surprised to see you state that ” As everyone is well aware at this point, IBM’s Backup Recovery Media Services (BRMS), while a fine program for modest use, is not particularly good as a backup tool for systems with more than 2 TB of capacity. It just stops working at an acceptable speed.” That’s utterly ridiculous and I suspect you know better and just misspoke. It appears you’re equating BRMS with limitations of saving to a single tape device (you can certainly scale by concurrently saving to multiple drives, but even that has some limitations). Key here is that BRMS is not the bottleneck, it’s all about the backup device(s) — and, of course, the native save/restore it depends on. That’s true for VTLs, too. Key benefit of a VTL is the dedup capability which enables practical/efficient networking between VTLs. Most important point you miss is the necessity of having a backup/recovery management product like BRMS even with a VTL as your backup device. It’s highly improbable that any large IBM i customers managing massive backups to VTL are using native save/restore and not BRMS.
Re: “It just stops working at an acceptable speed. ”
We backup ~15TB every 24hrs; where BRMS works fine for us as well; streaming from a v5000 to our VTL; where the volumes are then duped to physical tape libraries — often two or three at a time.
We find the throughput is “satisfactory” for the media involved in the process — because in the end; physical tape needs to move past the heads — and that ain’t fast — so our expectations are met.
In context: How fast do “most of us” need to get that tape in the hands of the Iron Mountain guy for his run next Wednesday AM anyhow?….