Microsoft toughens up its cloud
Welcome to Protocol | Enterprise, your comprehensive roundup of everything you need to know about the week in cloud and enterprise software. This Thursday: Microsoft gets in the zone, what tech employees told Protocol about their companies, and what happens when the geeks become the referees.
Also, don't miss our next Protocol | Enterprise event on Tuesday, March 23, at 9 a.m. PT: "Leading Through Digital Upheaval." Protocol's Joe Williams will interview Honeywell CEO Darius Adamczyk on how a 115-year-old company is making the shift to the SaaS era. Register here.
(Was this email forwarded to you? Sign up here to get it in your inbox every week.)
The Big Story
Microsoft's embrace of the cloud will go down as one of the most successful strategic shifts in software history, but it takes a long time and a lot of money to get the battleship pointed in the right direction.
That's why Wednesday's announcement that Microsoft will make availability zones standard for new cloud regions around the world is significant, and not just because the company suffered a widespread hours-long outage on Monday. It's a delayed recognition of modern cloud architecture by a company that built its first data center in 1989.
Here's why availability zones are important, and what you need to know about them.
- All cloud providers offer their services out of "regions" spread throughout the world. Microsoft actually offers the largest number of distinct regions across the Big Three.
- Availability zones are data center buildings, or collections of buildings, that are spaced out within a given region. Here's how AWS defines it: "AZs are physically separated by a meaningful distance, many kilometers, from any other AZ, although all are within 100 km (60 miles) of each other."
- Each availability zone has separate networking and electrical power facilities, with the goal of preventing something like the OVH incident a few weeks ago, where a fire caused by an electrical issue in one data center actually took out four data centers.
- Cloud customers have to design their applications to take advantage of multiple availability zones, but once they do, they significantly increase the resiliency of those apps.
Microsoft has been much slower than its rivals to roll out availability zones. That has contributed to the market's perception (backed up by actual incidents) that its cloud is more brittle than the ones operated by its rivals.
- Microsoft introduced its first availability zones within two cloud regions (Iowa and Paris) three years ago. They've been a standard part of AWS since 2008. Google was also early to offer zones.
- This week, after rolling out availability zones within 11 of its more than 60 cloud regions since 2018, Microsoft said that every country in which it operates will have at least one region that offers availability zones by the end of 2021.
- Going forward, all new Microsoft cloud regions will also have availability zones at launch, which should help bridge the cloud-services digital divide. Most of its current regions with availability zones are in rich countries.
Availability zones can't prevent all outages. Software configuration errors like the one behind Monday's Teams outage are a far more frequent culprit. But zones have been table stakes for big clouds for a long time, and rolling them out around the world will make Azure more competitive.
- It's also a reminder of how flush Microsoft is right now, on pace to do $66.8 billion in commercial cloud revenue during its current fiscal year.
- Microsoft spent $5.4 billion on capital expenditures last quarter, and CFO Amy Hood expects that number to increase during the current quarter. Zones are expensive, but that should easily cover the costs.
There's an irony here. All these new facilities should arrive just in time to protect cloud customers from a new period of intense, damaging weather caused by climate change linked to human activity, such as data centers.
A MESSAGE FROM INTEL
"We're moving faster now than we've ever moved, and we'll never move this slow again." ICYMI, catch a glimpse of what the future looks like for developers in this Protocol interview with Stacey Shulman, VP and General Manager in Intel's Internet of Things Group for Health, Life Sciences, and Emerging Technologies.
This Week On Protocol
Survey says: Check out Protocol's comprehensive look at what tech workers think about their industry's power, doing business in China and AI. One area of particular concern to enterprise companies: 44% of respondents think their companies should not work with law enforcement agencies.
It's totally automatic: Software keeps eating the world. Protocol's Joe Williams talked to Workato CEO Vijay Tella about how companies are trying to automate work applications, and where they are falling short.
Onramp to the cloud: Sometimes it seems like the Big Cloud providers are trying to own every enterprise service they can think of, but Fastly CEO Joshua Bixby thinks buyers are skeptical. As edge computing becomes more important, he's predicting (and Fastly is betting) that they'll want separate cloud and edge providers.
Five Questions For...
Adrian McDermott, President of Products, Zendesk
What was your first tech job?
I interned for an IT company for a year, where I wrote programs in COBOL. During the day-to-day grind of building software, I quickly realized a bigger lesson — we were creating products for people, which made me think about the importance of designing technology that's easy for people to use. While I did learn a few things about computer science, the most valuable lesson was actually about the importance of good customer service.
What was the first computer that made you realize the power of computing and connectivity?
I took my first steps in BASIC coding on the Sinclair ZX Spectrum. Many hours were spent loading 8-bit games from a cassette tape and playing them every day. It was revolutionary!
What was the biggest reason for the success of cloud computing over the past decade?
Data centers are a pain in the arse, and not where you should put your efforts. Zendesk used to run our own colocation data centers before switching to AWS, which also helped us establish a new kind of trust with customers.
Will the pandemic usher in a new era of remote working, or will we all come back together when it is safe to do so?
Remote working has plenty of benefits, such as tapping talent from outside the Bay Area's tech bubble and being able to onboard these new employees completely virtually. However, looking beyond the first day, companies need to continue fostering a culture of collaboration with teammates, building relationships with managers and connecting with leadership. In a remote world, these high-trust aspects need new rituals to sustain themselves and cannot rely on the catalyst of a shared space.
What is one book that changed your professional mindset?
"First, Break All The Rules" by Marcus Buckingham. This book uses data to teach the reader how to be a better manager. I recommend checking out the chapter that lists 12 questions to gauge employee engagement.
Around the Enterprise
- Microsoft released an easy-to-use software update for Exchange users. It's designed to mitigate the effects of the massive Exchange hack until customers can patch all their servers.
- Digital Ocean hopes to be worth around $4.8 billion when it goes public. The offering will be an interesting test of the public markets' appetite for boutique clouds.
- SMS-based two-factor authentication is bad. This report from Vice explains why.
- AMD is now shipping the third generation of its Epyc server processors. The product has generated a lot of interest in the data center amid Intel's stumbles.
- Sometimes the cloud jokes write themselves. Delta just took off on a three-year project to move off mainframes and into multiple clouds, which might produce turbulence.
- The environmental impact of cryptocurrency is real. Data is emerging that quantifies the societal costs of the massive amounts of computing needed for blockchain transactions.
- Docker raised another $23 million. The money will help fund its transition toward developer productivity.
- Remember that big Twitter hack from last summer? A Florida man pleaded guilty to compromising several high-profile accounts to facilitate a bitcoin scam and will spend three years in prison.
- Meanwhile, Twitter CTO Parag Agrawal told The Information that its decision to embrace cloud services has helped it move more quickly on the product-development front.
- Enterprise platform companies struggle with questions of when to kick paying customers off. Ben Thompson of Stratechery talked to leaders at Cloudflare, Google Cloud, Microsoft and Stripe about how they make those decisions. Spoiler: It depends on which infrastructure layer they serve.
Thanks for reading — see you Monday.