So, how AWS can leverage mainframe backups?

Rafa Suguihara
5 min readSep 18, 2021

“Mainframes are dead in the next years.”

Mainframes won’t die. Retail, Banking, Insurance, Government, and Airlines rely their operations on mainframes. And it won’t be different for the following years. Until those companies understand and unlock cloud superpowers to their systems, each tech professional’s responsibility is to integrate mainframes into new architectures and modern designs.

The primary motivators to modernize mainframes are:

  • skills
  • cost reduction
  • flexibility
  • agility
  • innovation
  • risk mitigation

All of those motivators are fair from a business perspective. Skill shortage is a fact. The appeal to learn Cobol is hugely smaller than others like Python, Java, and C#. Most of the professionals are aging, and new professionals want to know about modern platforms.

To meet the increasing demand around consumption on existing systems, MIPS needs to be the map to forecasts (believe me, it’s not cheap).

On the other hand, mainframes are solid blocks with fixed capacity that will lock money through less demanding year periods.

Accomplishing new business challenges on the mainframe is not easy, so the cloud is essential for this digital transformation. One example is this amazing 2018 re:invent session where HSBC presented a piece of work bonding mainframes and notification services through AWS using one of my favorite services (Kinesis). I was there, and still today, it’s impressive.

For those drivers, AWS present different strategies to modernize it. There are hardware emulation, a traditional lift and shift of workloads, indicated in mainframe shortage/decomission, middleware emulation taking advantage of API integrations, Automated refactoring using tools that will move Cobol to Java (automagically), repurchase to SaaS platforms and Manual rewrite. To know each one of those strategies in further detail, this great article brings more details of each.

Below is a graphic representation of what paths, average time and cost of each of the approaches.

That said, the mobilization for a modernizing journey is huge. The recommended approach is to create a business plan, but still a journey of some million dollars. To create confidence of C-level, there are some sweet spots to start, like mainframe backups.

A piece of experience: back in 2017, I was working for a large brokerage firm, and for audit reasons, we needed a backup of 4 years back. The Traditional Ops team requested the backup maintainer the tape with the backup. Then, the courier suffered an accident on an express route to our office, and a TRUCK ran over the motorcycle and destroyed the tape. The fine was high, and we weren’t able to present the data to regulators.

All companies need backups. They must be reliable, and the most important taking the narrative I gave: It must be accessible, reliable, and fast retrieval.

Beyond lines of code, we have databases. And those databases, even compressed, use a lot of space and time to be written.

Including costs must consider backup software, tape machines, the tapes or virtual appliances, and personnel to keep everything running. Long-term retention policies like accountant and tax (10 years) demand a high effort and different ways to maintain it. Some studies say that using LTO tapes and a good storage location is cheaper than moving to the cloud. But then we fall into agility, innovation, and risk mitigation. Not valid; try again.

In my research and experiences, I’ve reached Model9 (please sponsor me), an Israeli company with an exciting approach. Using zIIP mainframe processor, the agent compress and send data to AWS S3. AWS Direct Connect can make this connection due to data amounts or VPN (recommended for redundancy). All data will reside on S3, and its intelligence can make this burden incredibly cheap.

This approach can be the turning point for your company to pivot from mainframes to AWS. Costs will be reduced, and agility will increase. You’ll have 99.999999999% of durability and paying in the best case scenario 0.004 per GB / Month. Seems fair to use AWS and Model9 to lay your head on the pillow at night having all secured.

This approach will generate engagement from operations and certainly will draw the attention of C-level and financial teams to thrive the cloud journey, unleashing the most innovative ideas (remember that we can unlock a bunch of achievements).

References

--

--