Analysis IBM's Systems segment saw dismal revenues in its third quarter results, with declines in both servers and storage.
Z System mainframes are benefiting from the Watson-style cognitive computing efforts and both all-flash and software-defined storage are looking up, like beacons of light in the gloom. It's worth stressing that IBM's positive-looking software-defined storage wing – think SoftLayer and Cloud Object Storage – is separate from the glum Systems storage.
We think IBM is being driven by events in these areas rather than driving them, contrasting with its approach in the cognitive and cloud product and service areas. Although there is progress in both of these, it's not yet strong enough to offset IBM's revenue decline.
IBM CFO Martin Schroeter said in his prepared remarks on the third quarter revenue picture: "Turning to our Systems segment, there are some important market shifts in this business, like spinning disk to flash, the rising importance of the hyperscale data market, and new opportunities in blockchain."
"We're shifting our business, delivering innovation in our offerings, and introducing significant new capabilities. As always, our performance in the period is based on product cycle dynamics and portfolio transitions, and given where we are in the transitions in Power and storage, and in product cycles more broadly, our revenue and profit is down after a strong 2015."
The storage part of IBM's Systems business saw a 9 per cent year-on-year decline in revenues to, we calculate, $462m, meaning a $1.8bn run rate. Same old, same old, would be the quick judgement, but underneath the overall view there are beacons of light, with three growth areas bucking the trend. Schroeter's prepared remarks contained this info on storage:
Storage hardware was down 9 per cent this quarter, reflecting the ongoing shift in value towards software. Gross margin is down, reflecting both volume and price pressure.
The hardware decline was mainly driven by low-end and midrange traditional disk storage. Our high-end disk storage grew this quarter.
All-Flash Array revenue grew, as we have expanded All Flash technology across our product portfolio. We recently rolled out new products and transitioned to a full suite of flash offerings, making us competitively positioned.
And while not in our System segment, we also continued to see double-digit revenue growth in Software-Defined Storage.
The declines in low-end and mid-range traditional disk storage – think Storwize for example – were savage enough to send revenues down 9 per cent year-on-year. This despite growth in all-flash arrays such as all-flash Storwize and despite the somewhat surprising growth in high-end disk arrays. Eventually we could see a sustainable Systems storage business emerge, an all-flash array business probably.
IBM has yet to get properly to grips with hyper-converged systems, where it has no x86 server business to integrate with its storage, and would probably have to acquire, OEM or resell some other supplier's technology to have a presence. The idea of it providing a hyper-converged POWER server-based system seems a little far-fetched. A hyperconverged version of its converged infrastructure VersaStack reference architecture could be a possibility in storage boss Ed Walsh's mind, but this would seem a weak offering.
In the servers segment, meaning z Systems mainframes and Power, revenues were down year-on-year but were not separated out; so we have no specific numbers. We are told:
- The Systems segment overall saw revenues of $1.6bn, down 21 per cent annually.
- Within that, Operating Systems were down 11 per cent and hardware down 25 per cent.
- z Systems revenue was down 35 per cent y-o-y.
- Power servers revenue was down 31 per cent.
Schroeter said this about z Systems: "Our z Systems results reflect the product cycle dynamics. Seven quarters into the z13 cycle, revenue was down, while margins continued to expand. We continue to add new clients to the platform, and we are introducing new technologies like blockchain."
Big Blue is currently working with over 40 clients on pilot blockchain use cases running on z Systems.
The Power server side was less impressive:
Our Power performance reflects both our performance in a declining UNIX market as well as our growth in a growing Linux market. While our margins were relatively stable at the high end of Power, mid- and low-end margins were down, driving a decline in overall Power margins. We have been shifting our platform to address Linux, and in the third quarter, Linux grew at a double-digit rate, and faster than the market. It now comprises over 15 per cent of our Power revenue. Supporting that is our success with HANA, where we are bringing in new clients, and we are replicating this strategy with others.
With just 15 per cent of Power revenues coming from Linux there sure is a way to go.
Generally we don't get a sense of dynamism emanating from IBM in the server area, particularly with Power. It seems to react to trends more than drive them in that area. Its engineers must be looking at persistent memory ideas and NVMe over fabrics-style ultra fast access to shared storage. But we're not hearing anything about it, although the OpenCAPI initiative might reflect activity in these areas.
A market shift to hyperscale was identified by Schroeter but he said nothing about any IBM initiative in that area.
In general then, Systems seems an unexciting legacy business for IBM, which produces cash and helps fund the more interesting shifts to cognitive computing and the cloud. Or, as we might say with Watson in mind: "No shit, Sherlock." ®