Questions Arise Over IBM’s New Cloud Backup for IBM i
October 31, 2016 Alex Woodie
IBM raised some eyebrows with the recent release of its Cloud Storage Solutions for i, which enables IBM i shops to back up and archive DB2 and IFS data to its SoftLayer cloud. Much of the initial feedback to IBM’s first native cloud solution for IBM i was positive. But some IBM business partners are already questioning whether the solution will be actually useful for customers as it now stands.
Cloud Storage Solutions for i was a part of IBM’s big October 11 announcement for IBM i 7.3 Technology Refresh 1 (TR1) and 7.2 TR5. The software, which carries product number 5733-ICC and also supports IBM i 7.1, costs either $2,400 or $5,000, depending on how many LPARs you want to back up. You can read more about it in our October 17 story IBM i Finally Gets Native Cloud Backup.
The reaction from IBM business partner/competitor Dell EMC, which sells the DataDomain line of virtual tape library (VTL) appliances, was not exactly positive.
“Maybe we’re looking at it wrong, but I don’t see how you’re going to do a full backup or DIPL over an IP link from another site,” says Ed Fabiszewski, an IBM i practice manager with Dell EMC. “If you want to do full DR, you’re not going to do that over IP.”
To be sure, plenty of IBM i shops base their disaster recovery (DR) strategies on data sent across Internet Protocol (IP) or other protocols, like UDP, which is favored by applications that require low latency, such as VTL, WAN acceleration devices, and software and hardware-based high availability setups. However, in all cases, a secondary IBM i server must be present to accept the data.
While DataDomain is having success selling its line of high-performance VTLs into the IBM i installed base–the company has close to 1,400 IBM i customers in several years, Fabiszewski says–the company is clearly concerned that the IBM i cloud could chip away at this base.
Fabiszewski is right about doing an IPL over IP. While high-end VTLs like the DataDomain line can be used to initialize an IBM i server, that’s not really an option over an Internet connection. Considering that it takes 122 hours to send 500 GB over a 10 Mb/sec line, and 51 hours to send 1TB over a 51 Mb/sec line, IBM i customers should have other contingency plans to do a full restore, because it’s going to take a long time to do it over IP.
Even among other online backup providers who cater to IBM i, the typical solution calls for shipping a tape to the customer in the event of a disaster so they can recover their data. It’s not clear that IBM has the capability to recover a backup quickly and ship it out on a tape cartridge if the customers wants to recover their systems in less than a week.
Clearly, IBM isn’t providing a full cloud-based DR solution here. If you need DR, you can implement an HA solution, or rent space on a remote IBM i server living in the cloud. There are plenty of such services availed from the likes of Abacus Solutions, Baseline Data Services, and Connectria. If you’re a LUG member, you probably have a full HA setup, in addition to a full pencils-to-servers DR deal with IBM Business Continuity and Resiliency Service (BCRS) or Sungard AS.
Others in the IBM i DR business wonder about some of the fine print that has yet to come out regarding the IBM i cloud, such as whether it will support encrypted data, or whether the 1 TB data limit factors in compression.
“IBM is late to the game with a half-baked solution, in my humble opinion,” says the CEO of one IBM business partner, who asked that his name not be used. “It took them 10 years to come up with this?! IBM has not fully vetted this. They’re looking for guinea pigs.”
In IBM’s defense, Big Blue has been adamant that the new cloud backup offering will not be suitable for everybody. For starters, just look at that 1 TB cap on the amount of data a customer can back up to the cloud. That would barely cover the incremental weekly backup for many big IBM i customers, let alone a full system save of an entire database and associated system objects, which runs into the hundreds of terabytes for the biggest IBM i shops.
The IBM i cloud, as it now stands (or floats) is clearly aimed at smaller businesses, most of whom rely on tape to back up their data at the moment. As a tape-replacement solution for SMBs–not full DR–the IBM i cloud may live or die based on how quickly and easily data can be backed up and restored.
In a briefing with IT Jungle earlier this month, IBM i product offering manager Alison Butterill and IBM i business architect for application development and systems management Tim Rowe repeatedly stressed that the first release of Cloud Storage Solutions for i is just a starting point, and that it will evolve in the future.
Butterill told us the APIs were designed “so that you can send limited amounts of data up to cloud providers, whether it’s IBM or whether it’s others” (italics added). (The offering, you will recall, is being rolled out on SoftLayer first, but is expected to be available on Amazon Web Services, Google Cloud Compute, and even Microsoft Azure at some point in the future.)
Butterill added that IBM is looking to get feedback from customers regarding their use of this new product. She said there are “some limitations around file sizes and the speed of the links.” Her colleague Rowe said that getting the first backup into the cloud “might be difficult, depending on how large it is.” “This is the first step to provide the starting point for” building a full cloud solution, Rowe continued.
IBM is welcoming feedback from early adopters of the new IBM i cloud. It may get more than it bargained for.