Last year I described AWS re:Invent 2015 as being, in a word, “big”. While the 2016 rendition of Amazon Web Services’ marquee event was bigger still (32,000 attendees compared with about 17,00 last year), I think the operative word this year is “evolutionary”. The show had a notable absence of groundbreaking product releases, with energy focused on incremental product enhancements that address very specific problems. The idea that AWS has transitioned from the revolutionary to the evolutionary seems at a glance to be disappointing. From the perspective of their competitors, however, it should be absolutely terrifying.
Consider Andy Jassy’s opening remarks. He touched on a few very large and important numbers. Among them a $13B revenue run rate, 55% YoY growth, over 32,000 event attendees, and over 1,000 major enhancements. This represents a revenue run rate that is multiples ahead of the next closest competitor (Microsoft) at what is likely a comparable YoY growth rate. It’s impossible to make an apples-to-apples comparison because Microsoft does not breakout Azure from the rest of what it considers “cloud” (which includes Office 365 and…Xbox Live?). Based upon a reasonable interpretation of known variables, it’s safe to say AWS is significantly larger and growing rapidly by any measure. The law of large numbers doesn’t seem to apply to AWS at the moment.
No less important (if slightly problematic from a logistics perspective) were the 32,000 attendees that were present. At around $1500/ticket and several hundred per night for a hotel and being held the week after Thanksgiving, this is no small commitment on the part of attendees and their employers (thanks boss!). As important as scale is for the infrastructure and cost sides of the business, large-scale adoption and training on the part of technologists and their leadership is critical for continued growth in the space. Given the impracticality of learning and supporting multiple platforms, the adoption and training game is nearly zero sum and here, again, AWS has a sizable lead.
So why do these incremental improvements matter so much? Each is important in its own way as I’ll outline below. Taken in the aggregate, they represent a continued narrowing of opportunities to differentiate and compete with AWS. They also represent the advantages of scale and AWS’ laser-like focus on strategic execution.
1) New Instance Types
Table stakes in the cloud game, it’s been a few years since AWS announced any significant upgrades to compute families and sizes. Rolling out several new sizes of existing families and new options such as P2, flexible GPU and FPGA demonstrates a continued commitment to improving the real backbone of the platform. It’s Moore’s law in action.
From its inception, AWS has always been about enabling and empowering “the builders” – the developers and engineers who use the platform most directly. This is evident in the robust API’s and CLI behind all of the services, and the general configurability and ‘tweakability’ of everything on the platform. While this proves to be a great strategy that provides maximum flexibility, the downside to this approach is that seemingly mundane and standard tasks can at times become unduly burdensome to the novice user. Microsoft has taken a contrary approach with Azure, choosing to simplify things by wrapping many of them neatly for black box execution. This left customers with a legitimate choice between flexibility and ease of deployment. Lightsail begins to dampen that slight advantage for Microsoft, combining the flexibility and power of the underlying platform with a more straightforward interface for ease of use while preserving maximum flexibility.
3) Unleashing of AI and Deep Learning
Arguably the most consequential announcements this year were those pertaining to AI and Deep Learning as AWS seeks to “democratize” (Werner Vogel’s word) access to these capabilities. The advantages of scale here are tremendous as AWS is able to leverage the collective learnings and implementations of a massive user community. This should, in theory, greatly accelerate the improvements in these platforms. Look no further than the swag giveaway for evidence of how this will play out. AWS gave away an Echo Dot to every single attendee. That’s 32,000 people plugging away on the Alexa platform, building new capabilities, tying into new ecosystems, and providing valuable answers to the all important question – “did Alexa do what you expected?”
As Andy Jassy was prefacing the exabyte scale data problem and limitations of Snowball for solving it, I was joking with a coworker – “what are they gonna do, roll a friggin truck up to your datacenter?”. Well, turns out the joke is on me because that’s exactly what they’re going to do. Touche Andy Jassy…touche. This hard-drive-on-wheels is going to allow AWS an effective monopoly on some of the largest and most lucrative cloud customers – the ones with massive volumes of data. It will be interesting to see if Microsoft, Google, and Oracle simply cede this ground or come up with a similar solution of their own.
4) Green Grass
If Snowmobile is the “never underestimate the bandwidth of a truck full of hard drives” solution than Green Grass is the answer to the age old problem of “what do you do when the speed of light is too slow?”. Moving processing capability closer to the edge for IoT devices has a lot of potential ramifications in regions where internet access is restricted, slow, or inconsistent. Even in areas where internet connectivity is not an issue, sub millisecond latency could be a crucial capability in areas of manufacturing, advanced machinery (aircraft engines), agriculture, and of course, autonomous driving. It will be exciting to hear the 2017 IoT State of the Union and how customers leverage Green Grass.So what does this all mean? The adoption of cloud and the advancements that come with it are moving fast and furious; perhaps faster than even the most ardent optimists had envisioned. Look no further than the awkward “fireside” chat between Andy Jassy and Pat Gelsinger (CEO of VMWare) as they discussed the recently announced VMWare on AWS offering. This is a marriage that neither likely even thought possible 5 years ago. AWS is continuing to build on a well established lead and fully exploiting the key weaknesses of their deep-pocketed competitors while swallowing-up the smaller ones.
For Microsoft, the issue is trying to balance competing business interests against its cloud investments. Where the synergies are obvious (Office 365, Visual Studio Team Services), they are moving quickly. Where the interests are in conflict, we’re seeing more tentative hybrid approaches (Azure Stack). I expect this strategy to continue to play out as Microsoft focuses on defending its existing turf and carefully choosing its spots to do battle with AWS.
Google’s challenges are best summarized by a tweet I saw shortly after Google announced the reformation of the company under the Alphabet umbrella. One of the tech analysts following Google tweeted: “Google is focused…on everything.” Their lack of focus is in stark contrast to AWS and even Amazon writ-large. The ramifications of this focused-on-everything strategy are playing out as they continue to cede ground in the cloud to both Microsoft and AWS and perhaps even to Oracle.
I’m confident more will change over the next 12 months as customers continue their march to the cloud. I for one will be looking forward to helping them get there…and also to my daily dose of at least 3 new product enhancements!