Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.
This week: Redmond SaaS keeps failing, RISC-V is RISEing, and Meta is enforcing hybrid work.
1. Outage Outrage: What Can We Learn?
First up this week: The yo-yo nature of Microsoft 365 is causing great dissatisfaction. How can you avoid emulating the same mistakes?
Analysis: Poor update management
Microsoft’s communication about these “impacts” and “issues” is predictably jargon-filled and information-sparse. But, reading between the lines, it sounds like failed updates are to blame. Perhaps M365’s architecture isn’t up to the job, but Redmond’s DevOps team needs to do better testing and plan slower, more controlled canary rollouts.
Anirban Ghoshal: Another M365 outage
Microsoft … saw its SaaS-based Office software suite suffer an outage for the seventh time this year as users faced issues with Outlook Web, Teams, OneDrive for Business, and SharePoint for over eight hours. … Services were hit again just a few hours later [for] three hours.
About 18,000 users were affected at the peak. … The second outage saw over 5,000 users affected [by] the company’s seventh downtime in the last five months.
I count the eighth and ninth failures since then. @MSFT365Status describes #7, #8 and #9:
We’ve determined that impact … has reoccurred and are investigating the cause. … We’re seeing some service improvement, and we remain focused to investigate and address the source of the issue. … We’ve confirmed that service availability has returned to healthy levels.
We’re seeing a recurrence of the issue and a drop in service availability. … We’re seeing availability return to healthy levels.
We’ve identified that the impact has started again, and we’re applying further mitigation. … Our telemetry indicates that the service has remained largely stable; however, the incident is not fully resolved. … We continue to observe stable service health since we’ve applied our various preemptive mitigations.
And Rich 2 despairs:
We have migrated to all things MS cloudy and Teamsy over the last month or two where I work. The explanatory email that went around from the IT department to everyone explained that this was because it was (and I’m not joking here) “best in class.”
Oh the temptation to reply with a company-wide email correcting this utterly misplaced thinking… Thankfully, I only have to use Outlook and Teams (OM*G!!!! What a heap of utter ****) so am probably spared the rest of the cloudy goodness. And quite frankly, I look forward to an outage to give myself some peace from Team’s incessant pings and popups and general ******-ness.
2. RISE to Take RISC-V to Next Level
The RISC-V Software Ecosystem (RISE) project sets out to build all the software and firmware infrastructure needed to make RISC-V chips a success. How long before RISC-V is a viable server platform?
Analysis: Juggernaut gains momentum
ARM should be worried: Five of its biggest licensees are founding members. Smart DevOps people will feel the wind changing and prepare for a world where infrastructure runs on RISC-V too—not just x86 and ARM.
Mike Vizard: RISE Open Source Project
“Eliminate duplicate work”
The Linux Foundation Europe, in collaboration with RISC-V International, announced [the] RISE project. … Multiple organizations will contribute engineering and financial resources to harden critical open source software for use in commercial applications and enterprise IT environments. Founding members … include Andes, Google, Intel, Imagination Technologies, MediaTek, Nvidia, Qualcomm Technologies, Red Hat, Rivos, Samsung, SiFive, T-Head and Ventana.
The overall goal is to use a more holistic approach to eliminate duplicate work that each member of the RISC-V community would otherwise do separately to create an application processor … (includes software development tools, virtualization support, language runtimes, Linux distribution integration and system firmware). … Project deliverables will be prioritized by the RISE Technical Steering Committee (TSC) as the project looks to address performance, security and reliability issues.
ELI5? ffkom translates:
Corporations fed up with paying the “ARM instruction-set tax” aim at establishing some competition. (Which is kind of ironic, given how much the same corporations totally want monopolies in their own respective businesses.)
With an alternative angle, here’s Scott Foster:
The RISC-V Foundation was established in Delaware in 2015. [But] in 2020, the Foundation was incorporated in Switzerland as the RISC-V International Association … to avoid potential disruption caused by then-president Donald Trump’s anti-China trade policy. … The launch of [RISE] makes open-standard integrated circuit design and open-source software even more of a challenge for the US government’s efforts to stop the development of Chinese high-tech and bend Europe to its geopolitical will.
Proprietary instruction-set architectures from Arm are seen as high-risk by the Chinese due to potential US influence on their owner. … China has involved itself in RISC-V from the beginning and that – particularly in view of the Biden administration’s liberal and expanding use of sanctions – has turned out to be a very good idea.
3. Meta Tells Everyone to Come Back to the Office
Unless your job is “officially” WFH, you’ll soon need to at least be hybrid. So say Meta’s HR zuckerdroids.
Analysis: Dumb move
Despite recent high-profile layoffs, there are still good DevOps jobs to be had—if you’re good at your job. Thoughtless moves such as this will see many of those good staffers quit, the remainder being average-to-poor achievers.
Ginger Adams Otis: Meta Requires Office Workers to Return to Desks
“It’s time to come back”
Roughly 75% of the company workforce … will have to start coming in three days a week … as the company shifts to a more structured hybrid schedule. … The change is meant to foster “the collaboration, relationships and culture necessary for employees to do their best work.” … It will take effect Sept. 5.
Other tech companies, including Alphabet’s Google, Apple and Microsoft, have also told workers it’s time to come back. … In a 2022 survey by ADP Research Institute, two-thirds of the workforce said they would find a new job if required to return to the office.
Very bad idea, predicts gweihir:
So they want to lose anybody competent? Because that will be happening. Anybody competent can get an online or at least mostly-WFH job these days, and the employers offering that are the only ones that get enough applications from competent people.
[The rest] go to the office not to produce results but to socialize. I can understand [they] feel very threatened by WFH. With more WFH, it will become very obvious who is what.
Incidentally, isn’t this the same “Meta” that wants people to work in the “Metaverse”? Talk about an abysmally stupid move.
With a more nuanced view, it’s Peter2:
This would depend on what the job is. Programming? Yes, working from home would tend to be rather more productive as morons can’t wander over and disturb you.
Anything involving communication? Sitting around a table with some paper to scribble on is pretty indisputably the golden standard that other solutions are tested against.
How does this keep happening? Here’s metadat with an explanatio: [You’re fired—Ed.]
One school of managerial reasoning is nobody is that critically essential: Most business problems can—eventually—be solved by throwing more money and bodies into the pit until there is a breakthrough. I don’t like it, but when pressed … many execs will cave to this rather than fall on their own sword and fight against someone like Zuckerberg.