Broadcom – a mainframe charm offensive - And I found it pretty convincing

Broadcom has just taken a group of analysts and journalists to Boston, to imbue us with a sense of the seriousness with which it takes the Mainframe – and with keeping its mainframe customers happy and productive.

On the whole, I think they succeeded. I met quite a few people from its Mainframe software division who I already knew from IBM and Computer Associates – it certainly has plenty of mainframe experience and expertise at the top. Greg Lotko (Senior Vice President and General Manager for Broadcom’s Mainframe Software Division) gave an impassioned defence of the mainframe and its continuing importance to the very largest companies. Even if you do plan to get off the mainframe, he says, and he isn’t convinced that that is going to be possible, for many players, this is going to take time (years) and you can’t afford to neglect the management of a mainframe resource critical to large enterprise operations whilst migration continues. And, if there is a risk that migration off the mainframe never actually completes, neglecting mainframe management and optimisation in the meantime represents significant unmanaged risk. The Broadcom Mainframe Software Division (and remember that Broadcom divisions have considerable autonomy, in practice) believes strongly in running the right workloads on the most appropriate platforms, and that large enterprises building hybrid solutions, with a significant place for mainframes, for the foreseeable future.

I like the Broadcom mainframe software division, and I liked the people I met, and their palpable enthusiasm. Given its separation between divisions I can’t speak for the rest of Broadcom, but Bloor’s Justin Speake (who, in an earlier life, worked for Computer Associates, now part of Broadcom) sees Broadcom’s strategy as “to maximise share of wallet long term from the world’s biggest companies as this keeps its price-to-earnings (P/E) ratio up.” – which makes me believe that it really is interested in long-term customer relationships. This also means that if the Mainframe division is successful, other divisions should learn from it. What it doesn’t mean is that Broadcom is particularly interested in small companies – fewer larger customers are easier to manage – or that it is interested in simply maintaining the status quo instead of growing its relationships – all of which implies changes for existing customers. Forced change, after merger/acquisition especially, can be uncomfortable and customers will have to be persuaded that change is exciting and is delivering improvements for the customer relationship, as well as for throughput and resilience. The CA acquisition (where most of Broadcom’s mainframe technology originated) has had time to mature; the VMWare acquisition (say), in a largely independent division, which is featuring in the press currently, hasn’t had a chance to mature yet.

So, let’s dig into some of the things which I liked about Broadcom’s mainframe offerings and approach. I’ve already mentioned its belief in running workloads on the most appropriate platform and the concomitant focus on hybrid architectures. Some workloads will move but the mainframe ain’t going away anytime soon.

The report from Rubin Worldwide titled The Economics of Technology Investments goes into more detail on balancing risks and returns in order to achieve an optimised hybrid environment, incorporating both existing and new investments and leading to the best business outcomes. Some of its insights may surprise you. For instance, Public Cloud spending (AWS, Azure, Google Cloud, IBM) hasn’t even reached 10% of the $8 trillion annual worldwide IT spend. It says that “compared to average performing organisations, the absolute best-in-class organisations use an overall IT mix with more public cloud (10% more), more mainframe (10% more) and less distributed (20% less) compute power”. Its conclusion is that, despite the popularity of public cloud migrations in the press, “the vast majority of compute power – 85% – is still on-premise. It is the distributed technology asset class that is losing ground to the cloud; the mainframe asset class is steadily growing”. Which means that it might be an idea to spread mainframe expertise within your organisation.

This seems to be one of the “value adds” from Broadcom, as communicated by Vikas Sinha, VP of Customer Experience and Alliances. They provide training to improve mainframe expertise, they will find you expertise if you need it, and they will let you hire its mainframe experts for your own organisation, without charging a finder’s fee. Broadcom is going “beyond the code” with its mainframe offerings. You need more than just good applications and tools, you need people and business skills – and business vision in a tech partner.

So, for example, you can now find at Broadcom:

Jeff Henry, VP of Strategy, Product Management and Design, talked to us about amplifying mainframe value. This is a good point – cost is easy to measure, value isn’t, but given sufficient value delivered, cost is almost irrelevant. These days, a lot of the value from automation of large business systems is already being harvested (with, in many cases, incumbent mainframe systems), so innovation may have less value to give, and cost becomes important, but we should always balance cost and value so as to maximise business effectiveness, not chase cost reduction alone (which can impact service delivery). Henry told us about the place of Broadcom’s Open Solutions in DevOps, AIOps, to shift Ops from being reactive to proactive (WatchTower), Data Management and Cybersecurity/compliance, in amplifying mainframe value. Some of these currently seem a bit more Open than others, to my mind, but I did like the WatchTower observability story.

Perhaps the most significant news from this workshop was around Broadcom’s adoption of Commercial Open Source solutions. Rose Sakach, Offering Manager, Open Mainframe and Open Source went into its open approach to Mainframe Modernisation, with the support of some practical use cases. I haven’t got room to go into the Open Mainframe Program (OMP) [https://mainframe.broadcom.com/blog/powering-up-open-mainframe-ecosystem] and Zowe [https://www.broadcom.com/products/mainframe/product-portfolio/open-source-mainframe-zowe] here, but my impression is that Broadcom is very serious about this, and about contributing back into Open Source communities. There is a bit of a cultural risk here and big companies do not always integrate successfully with Open Source communities but I haven’t noted any particular problems yet. Perhaps some of the impetus for this is that Broadcom actively wants to collaborate with other mainframe vendors. The mainframe is not a platform that can afford to have vendors selfishly squabbling over proprietary solutions, it already has image problems of its own.

Broadcom’s embracing of open technology isn’t just about Open Mainframe, and having an API mediation layer, and developing an API State Monitor for Zowe – and putting stuff on git. It seems to want to open up Broadcom to Open Source DevOps products and make Broadcom’s UIs as attractive as possible to Open Source developers. As I said, I think Broadcom is very serious about this; but we may have to wait and see how Open Source communities as a whole take to it.

Broadcom’s charm offensive on its Mainframe support was pretty convincing (as you’d expect; Broadcom wanted to talk to us and invested a lot in this analyst workshop). What could go wrong? Well, assuming that Broadcom can deliver on its promises (and I see no reason why it shouldn’t, it has the expertise), I see reputation risk, cultural risk, and political risk. It’s “beyond the code” offerings won’t help if people don’t know about them (so good messaging, reaching beyond the mainframe community, will be important) or even simply refuse to engage.

The mainframe culture places resilience and maintaining services above (not in place of) innovation; other IT cultures may be different. Mainframe customers have long memories, and they can remember when the mainframe business was seen as a “cash cow”, with locked-in customers that could be exploited. The combination of a somewhat conservative viewpoint and a distrust of vendor altruism will not help sell the mainframe platform to the cloud zealots in a company. Broadcom needs to rebuild trust (as it seems to be working on); conversely, its customers need to invest in cultural change and the elimination of technology silos. Initiatives such as Zowe, Open Mainframe, and mainframe DevOps, will help to open up the mainframe environment more widely.

At Bloor, we think that mainframe data, and the business knowledge locked up in mainframe systems that have been operating effectively for years, is of immense value to an organisation and shouldn’t just be discarded. Mainframe modernisation – running the right workloads on the most appropriate platforms – is the best story for businesses. We just wish that the term “mainframe” would go away and be replaced by something like “Enterprise Server 3”, so that we can just carry on using what is, these days, a powerful, effective and highly resilient technology, where it makes business sense, without worrying about any last-century baggage that may be associated with it.