Comments on: IBM and the resurrection of the mainframe

Steve Mills runs IBM's $20 billion software business, but don't get the idea that he has his head in the cloud. The 34-year IBM veteran doesn't put a lot of stock in the cloud computing for his enterprise customer base.

[dlcusa]

It's MVS, not VMS (that was DEC's operating system).

[Rusty Digital Marketing]

Mainframes and VM (IBM's Virtual Machine OS which hosts Linux) are definitely regaining their relevance to medium and large businesses.

Mainframes are clearly the most efficient and scalable form of computing, from a hardware cost, an environmentals / power viewpoint, and from a systems management perspective.

The only people who think it makes sense to have huge server farms of bewildering complexity are Microsoft and possibly Oracle (who is trying to elbow its way into the operating system market with its clustering technology). Oh... And all the techies who think that Microsoft certification is their key to long term prosperity.

[sal-magnone]

You mean second resurrection of the mainframe, no?

[Dana Gardner]

Great piece, Dan, we'll have to call you maestro now. ... Mills and IBM faces a massive trend of lowering total IT spends for ever-larger sets of apps and user organizations.

If IBM draws revenue from service providers operating at high utilization instead of lots of enterprises running on low utitl, the addressable market is impacted. More bang for IT buck leads to fewer wasted bucks that end up on IBM's bottom line.

SPs will demand more value, less margin for the highly competitive supplier base. IBM's margins under pressure. IBM begins looking like Sun five years ago. Race to commodity on the picks and shovels to the cloud providers.

This is not science fiction, it's economics.

[daver]

Mainframe apps presuppose user cohesion and software conformity. I see neither dynamic happening in the future. There are some things that just can't be simplified or reduced to a common application. They can only be addressed by a huge multitude of multitenant apps served on many different servers designed specifically for each app's needs. Yes, it's messy, but there's no way around it. While I see a place for what IBM is doing - especially in the market of enormous multinational organizations - it is ultimately a niche market.

[connieg10]

The interesting thing about the mainframe is that it is really designed for multi-purpose computing, so while it won't be the best platform for each individual app or service that will comprise a cloud of SaaS offerings, on the whole it will serve as a SaaS utility quite nicely. S. Mills quite accureately stated that most applications are not multi-tenant today, and the solution to making them multi-tenant is to virtualize them., rather than to rearchitect for multi-tenancy and show up late to the market in 2 years. The mainframe provides massive virtualization and is capable of sharing resources from the processors, to the I/O, to the network, etc. When you consider that most SaaS application offerings will run at some small utilization rate, even if the mainframe is not an optimal platform for hosting the single application, IT IS the optimal platform for hosting them ALL, and dynamically allocating resources based on service level agreements. With Linux and the open stack, plus huge strengths in data management, you can litterally run an entire cloud in the mainframe, making fairly complex cloud computing a lot easier from the provider perspective, and still invisible technology to the end user. The mainframe has broad applicability in the Cloud market which is focused on serving volumes of users. No platform does volume serving better than the mainframe.