Follow-Up: vCloud Director SP: The Need for UI Improvement!
Its been a couple of weeks since I wrote this post on the need for improvement with the vCloud Director UI and the response I’ve had to the article through the comments section, on LinkedIn and Twitter and also in the vCloud Director SP v8.0 Beta Discussion forums has been positive and supportive. With VMworld fast approaching it’s going to be interesting to see if there are any announcements around the future vCD SP and if there are indeed any movements on the need to have some form of improved UI…as shown below the Poll I conducted on that blog post reflects the reality that there are only a small percentage of vCloud Air Network Partners capable of using the new features in the vCD SP Editions.
rob.purdySubmitted on 2015/08/21 at 04:54
I’ve put this together from 100’s of meetings with VMware both EBCs, World, SP Councils etc. Some of it is supposition and VMware wouldn’t confirm for me… so it may not be 100% accurate but here goes. Also I’m not sure how much I can say given I’ve spoken to some pretty senior people at VMware plus I’ve been privileged to sit on their Service Provider Advisory board where there are a few things discussed that they put under NDA. So I’ll try to make a best judgement around anything that follows and I’ll try to put it in a timeline perspective…
0) vCloud was created because internally, VMware couldn’t modify vCenter fast enough to add features for cloud like adoption. So the decision was to spin out a dev team, let them go as quickly as possible and use the vCenter APIs to develop something that would allow SPs and the Enterprise (that was the real goal) to have a private cloud and stop the noise of AWS/Cloud that was humming in the market….BUT:
1) It was pretty much doomed from the start when they first released v1. It was terrible and that’s being nice. Poor vSphere integration, lacked features vCenter had. So any internal sysadmin didn’t want to go near it…remember it used to use GUIDs to identify VMs so you had to look for mappings? You couldn’t back it up or restore it… and the effort you had to (and arguably still) do to integrate it and make it work well didn’t justify the return… poor start team. Why add another 100 hours to the build of your VMware environment to get something that limits you? most decision makers saw this…it had promise… BUT:
2) They wrote it in the wrong language – Flash. They picked flash because it was quick/easy/looked nice but almost as soon as VCD 1.1/2 was released, (from my hazy memory), Adobe announced they were stopping investment. The world turned almost overnight to HTML5 and that meant a big re-write for them. Flash is horrible, zero mobile support (remember Steve Jobs), Flash is terrible, we all know it, just accept it. But they couldn’t justify the rewrite cost (and time delay) because…
3) Most customers weren’t deploying it. Sure there are some big installations around the world, mostly larger customers that had internal development shops and multiple business units, but the trouble was the only real adoption they were getting was with the service providers and they were getting it for free through the VSPP. At this time, VMware was trying to match up the SP’s and Solution providers and they needed the SP’s to have a standard platform with APIs etc… They were also backing a number of Telcos (remember vCloud Data Center partners) to try and lift them up to the emerging AWS, MS, (looming Google) threat of public cloud. So giving it away for free was a good adoption strategy. But unfortunately the vCloud Data Center program failed… I cannot go into why… Also for enterprises, Dev shops weren’t adopting it even though Lab Manager was dead because VCD didn’t have all the features of Lab Manager…
4) It was/is still a complex interface – it didn’t flow well and ultimately that’s why VMware themselves, when they launched their ‘VMware Hybrid Cloud’ trial platform in the US, didn’t use their own interface! They realised it was too clumsy and complex so they wrote something simpler for the masses. One insider told me it was at that point the interface was doomed well before it was announced… so its demise has been on the cards for years.
5) So now at this point VMware have a flash (dead) interface, very small enterprise adoption and mostly it doesn’t have the enterprise features, they aren’t even using it themselves and SPs are really the only ones using it but there’s no revenue to fund the unit because they gave it away in the VSPP. So the decision is, buy another company that has a portal (vCAC/vRealise) and shift VCD to the SP division because if we kill it all together we might lose the SPs totally, particularly the big ones that integrated it to their own portals (see next point). But at the same time, VMware realised that they could not viably get the SPs to compete with AWS/Azure so they had to do something themselves, hence vCloud Air was born. (just go look at the timing…think late 2012/early 2013)
6) So in 2013, most of the SPs using the VCD interface are those that cannot afford to develop their own or buy something – but that’s the trick, those that can afford to develop/buy, make up a significant % of the VSPP revenue. So say the top 20% of VSPP SPs represent 80% of the revenue (I don’t know the exact numbers) and of the top 20%, 80% of those SP’s have their own portal because they are telcos etc that have merged VCD APIs with existing portals or written their own. So for the VMware SP business unit that own owns a non-revenue generating platform, facing a full rewrite to shift away from flash and make it more usable again doesn’t justify the return. Also those top revenue SPs are the ones that get a voice at the highest levels in VMware and influence the strategy, not the 1000’s of SPs on 3600 plans or less commenting in forums unfortunately. So those that have the most influence, don’t need an interface (i’m generalizing a bit)… and even if they used VCD extensively, they can probably afford to license something and replace it because VDC isn’t their main line of business (telco lines, outsourcing etc is)…
7) So in 2014? VMware turn to the partners (ISVs) and suggest to the smaller SPs, go talk to some of our partners and buy their product. Flip the strategy and push it as an API layer to provide some standards to the SP community, keep very close tabs on the midlevel VSPP subscriptions and if you feel they are faltering, make sure you position vCloud Air to migrate, because if they use the vCloud APIs and connector, it should be pretty easy to get them into vCloud Air… pretty simple really…
So I’d be surprised if they resurrect it – and if they do it shows that their strategy is all over the show. Its alive, its dead, its alive etc… Just remember, SP revenue makes up less than 10% of VMware’s… so it’s a pretty small voice. As I said in my tweet last night (and I’ve made a few calls internally to VMware contacts in the last six months) there is sure to be an announcement this VMworld, but no one will say which way… they are being very tight lipped about it.
I asked Rob if I could repost this and he agreed because we share the same passion for VMware and vCloud Director and we both work for SPs with significant investments in the platform… The question has been asked of me since the original post went up about the possibility of an enhanced UI taking away the advantage SPs like Zettagrid, Datacom or iLand have by being able to develop against the vCD SP APIs but in reality a better UI that all vCloud Air Network Partners can use can only serve to strengthen the Network…and in turn that helps VMware compete against the likes of AWS, Azure and other IaaS Providers.
There is a strong group of vCD Supporters who hope the news coming out of VMworld next week is positive…time will tell!