VCE's new VxRail hyperconverged appliance sold 146 units in its first 44 days on sale and the EMC server unit's new leader Chad Sakac says that impressive start means he's betting the outfit will become the leading hyperconverged systems vendor by the end of 2016.
Nutanix is felt to be the revenue leader in the field, having recently revealed a $100m-a-quarter run rate. VMware has revealed about $100m a year of revenue just for its VSAN product and claims revenue leadership for hyperconverged software on that basis.
Sakac is confident VxRail will take VCE past both figures handily before too much time has passed. Firstly, sales tend to accelerate once a product hits the market. Secondly, Sakac told The Register yesterday that Dell has started re-selling VxRail which should speed things up too.
Over time he expects Dell will replace Quanta as the supplier of servers inside VxRail rigs. It makes sense for VCE's soon-to-be-owner to do so because Dell makes cracking rack-mount servers and is a market leader. Dell's also mastered the dark arts of figuring out just how many servers it will need to make on Tuesday afternoons in May and holding no more stock than is necessary to meet that demand. The likes of Quanta tend not to get out of bed until asked to make a few thousand boxes. Once Dell owns EMC, and therefore VCE, Sakac thinks supply chain and manufacturing improvements will drive VxRail costs down and make it even more interesting to buyers.
But Sakac said Cisco will remain VCE's choice in its larger vBlock offerings. Blades and chassis-scale kit, he said, is the natural building block for vBlocks. While Dell has a fine product, Cisco leads in blades. Cisco also likes the channel VCE provides, not least because it represents a decent chunk of overall UCS sales. Most importantly, customers like Cisco inside vBlocks. Dell sees no reason to imperil VCE's $3bn annual run rate, so Cisco stays.
Sakac also revealed that VxRail uses only about 20 per cent of code derived from VMware's failed EVO:RAIL hyperconverged software offering. That's not because that software was bad, but rather because the field is moving very quickly.
He also explained why, in his view, EVO:RAIL did not work. Sakac's analysis is that VMware got the licensing and pricing wrong and miscalculated the scale at which customers would run the software. Customer feedback about those messes, he said, was delivered to the partners who were trying to sell EVO:RAIL. By the time VMware heard and responded to criticism, it was too late to turn things around. While VMware has of course sold through partners for years, its ecosystem was not well-equipped to sell a platform product.
VCE, Sakac feels, was born to be a platform company and so has an easier time devising and explaining how software adds value to a hardware platform.
Which he would say, of course, now that he leads VCE. Although Sakac also acknowledged that the hyperconverged market is getting both busy and crowded … and that's before Microsoft makes a proper entry with Azure Stack and associated appliances. EMC and Microsoft get on very well, he said, as the former has provided SANS-a-plenty to underpin plenty of the latter's Exchange and SharePoint. But the two never really had a strategic relationship. Sakac said expects Dell's very close relationship with Microsoft to bring EMC and Redmond closer. ®