This is part of a series on SNIA BlogFest2
The second session was hotly anticipated in the bus on the way over. After such a strong showing by DiData, I was hoping for a day full of focus on customers and their issues, and how the vendors were addressing them with cool technology or solutions.
I feel I should disclose that I have worked for HP, as a contractor, back in 2000/2001 as part of their APIT division. I was also a bit of a fan of HP-UX, and quite liked Medusa and MC/ServiceGuard, but I have never owned an HP calculator. The HP CROs and spectrum analysers were always the best gear when I did electrical engineering at uni.
I am troubled by this session, and what to write about it. So I shall begin with the positives, and then try to explain my concerns as best I can.
The Kickoff
We were greeted by Orsi Sulyak (Storage Marketing Manager), Mark Nielsen (StorageWorks Business Manager) and Jason Davis.
The kickoff was quite good. Mark gave us a general overview of HP’s current thinking on the market, and echoed many of the same themes that DiData had: converged infrastructure, virtualisation, the need for operational cost savings.
Mark stated that about 70% of customer spend was on “just keeping the lights on”, and only 30% on innovation. Their customers are also suffering from IT sprawl and the breakneck pace of change.
Power and cooling were called out as a specific part of the HP’s approach, so much so that they have marketing phrases to match: “Adaptive Smart Cooling” is apparently a feature of some of their gear, which my notes tell me relates to “smart PDUs” you can install in your racks/kit which dynamically adjust the power draw to match the demands of the kit. That’s quite useful when power consumption keep climbing, along with the price of electrons. Containing power costs is a clear customer concern at the moment.
Lots of customers continue to use EVA storage, and HP is keen to keep them onside, so there’s no abandoning of that product line on the radar.
In fact, HP have essentially four product lines catering for different tiers of customer need when it comes to storage: 3PAR, EVA, Virtual Tape, and p6000. I asked if customers clearly understood which product was for which use case and Mark assured me that HP were confident of their product positioning. I’m not sure if that has translated into actual customer awareness, but I didn’t press the point.
Sadly, that was the last we heard about anything much to do with converged infrastructure. Mark handed off to Jason who launched into a slide deck on 3PAR.
All 3PAR, All the Time
This is where things started to go a little awry.
I’m reliably informed by Rodney that 3PAR gear is actually quite neat, and there are a lot of good stories to tell about it. Unfortunately, we didn’t hear about them because Jason was intent on ploughing through what turned out to be a pretty generic sales slide deck.
I tweeted extensively on the technical details of 3PAR presented on the day, so I’m not going to repeat myself here. If you’re interested, you can download a copy of the slide deck here [.pptx download link].
Instead, I want to talk about how the session was run.
After starting the session talking about convergence, HP spent an hour and a half talking about 3PAR. The first slide mentioned four product lines they wanted to get through, which was always going to be tricky, but to then concentrate on the one product line was puzzling.
And the type of concentration was odd. Jason gave us a technical deep dive, but insisted on spending quite a bit of time on common storage terms with an audience of bloggers with considerable storage experience. We know what RAID is. We know what dedupe is. We know what snapshots are. It’s beyond me why HP didn’t quickly mention the context (e.g.: it does de-dupe) and then tell us why 3PAR is special. More importantly, tell us why this specialness is important to customers.
For example, we spent quite a bit of time on the fact that 3PAR uses custom ASICs, but the stated goal for HP is to use 80% commodity components. I have no idea what these two things mean in combination. It felt a lot like HP were saying “hey, this thing is full of neat, shiny technology!” Umm.. ok. So?
HP were well aware of the time limits for the day, and that we were on a tight schedule. It was their slide deck, and these were all seasoned sales/marketing people. I’m baffled as to why they didn’t manage to run a tighter presentation. As it was, we ran into lunch, which was only meant to be 20 minutes long as it was. I was hoping for a bit of social chat over a sandwich, but no dice.
We got sidetracked talking about copper backplane interconnects, and what precisely a Common Provisioning Group is. When you have to spend a lot of time explaining what your marketing jargon means, that’s not a good sign.
One incident highlights what was, for me, the big problem with HP’s session. Near the end of the 3PAR portion, Mark, who had been silent the whole time Jason was giving his spiel, asked us what we’d like to talk about. But before we could answer, Jason jumped in to say that there were just a couple more things he’d like to cover on 3PAR. I just shrugged and said “well, it’s your presentation” and let them continue to run out the clock.
It’s a real shame, because HP missed out on an opportunity to connect to the local community. The 3PAR product is new to Australia, and there’s not a lot of local experience, but plenty of interest. To get a generic, and somewhat dull, sales presentation was disappointing.
Lessons To Learn
There are several lessons here for both vendors and the bloggers.
As bloggers, we need to more clearly define the objectives of the day, and provide better guidance to the vendors. However, we also need to educate the vendors on how to handle a blogfest, because there are vastly different capabilities from the teams in Australia.
The session from DiData covered a lot more ground in less time than HP. The session from Dell also managed to cover a lot in little time, and was interesting and engaging, even while providing some detailed information about products. More on that in the next post, but it highlights the differences in approach taken by the different vendors.
To that end, rather than dive into what I think would have worked better, I’m going to put together a separate post at the end of this series to provide guidance for vendors on how to deal with a blogfest more effectively. It’ll be based on the sessions from both SNIA BlogFests, and tips from some other sources as well.
Hopefully it will help both bloggers and vendors get more out of their sessions in future.
More Good Stuff
So that I don’t end on a down note, I want to highlight some good stuff that came out of the session, despite its flaws.
3PAR can dynamically change the type of RAID protection for LUNs, online. It can also do sub-LUN tiering, online. There is some capability to do this “automatically”, using an additional management tool/software, but manual movement is still good. Being able to move workloads around online is pretty much mandatory when you’re talking about “always on” infrastructure. Having to take an outage for what is now routine maintenance is no longer acceptable, so it’s good to see these features becoming more prevalent.
Remote replication can be done over IP or FibreChannel, and if you use IP, there’s a dedicated port for it, which is good to see.
3PAR devices are also manageable over IP, unlike EVAs which still require a FibreChannel connected server to manage them, though at least now you can virtualise the management server. 3PAR also has a vSphere plugin for management, and good support for VMware integration.
Get Thin, Stay Thin
One of the standout messages from HP was the “It’s easy to get thin, but staying thin is harder”, which relates to the challenges of thin-provisioning. This message was echoed by other vendors on the day, and it’s a good one to get out there.
Turning on thin-provisioning will likely give you a massive improvement in disk utilisation, which looks fantastic on marketing materials. You get to claim your storage “saved” 80% or something with some workloads. This is only sort of true.
The reason the technology works so well is because your customer was doing an incredibly poor job of managing disk utilisation. Honestly, claiming that you’re now making a lot more money because you’ve started actually invoicing customers isn’t something I’d be crowing about too loudly.
In Summary
Overall, I was disappointed in HP. This is not a small company, and I felt they could have done a better job. I was expecting them to, in fact.
I take no pleasure in saying what I feel are fairly harsh things. Quite the opposite. I would have preferred to write about how this neat technology I knew little about was doing some clever things that had obvious value to customers. I’m trying to be gentle and fair, but firm. I wanted to like their stuff, but I’m left feeling rather ambivalent.
Despite the mammoth amount of time we spent discussing it, I’m not at all clear on why 3PAR might be better for customers than competing devices from other vendors. I hope the marketing gets sorted out, and soon.
Better luck next time, hopefully.
Merch disclosure:
- Lunch: Some rather sad, limp looking sandwiches cut into soldiers(?!) that were actually fairly tasty, and some mini-hamburgers that were nice, but difficult to eat. Various softdrinks, though I stuck to water after too much coffee at breakfast.
- A wireless USB mouse. Cheap, plastic-y construction, but it seems to function ok.