Factoids about the Cloud Battles
March 10, 2020
DarkCyber noted “Stress Test the Cloud: Alibaba Cloud, AWS, Azure, GCP.” The write up presents “factoids” and observations based on these factoids in a helpful way. Here are the points which captured DarkCyber’s attention:
The cloud will be the way of the future in computing. The meltdown of Robinhood’s trading platform was pegged on stress. When a cloud system is stressed, it may and will fail.
Amazon Web Services
- “Amazon’s e-commerce business is the market leader in the U.S., Europe, and close to number 1 in India”
- “AWS is very much battle-tested and constantly “stressed out” by its parent company’s core e-commerce operation. It has moved all of its businesses onto AWS, and off of other systems like Oracle, after a multi-year effort.”
- Amazon’s businesses are generally not prone to unexpected spikes in traffic, which happens more to social networks like Facebook, Twitter, and Weibo.”
Amazon’s system may not be optimal for surprise spikes.
Alibaba Cloud
- “Alibaba’s core e-commerce business has many similarities to Amazon’s…”
- “This accomplishment is well-deserved; Alibaba has basically created and survived the mother of all stress tests.” The reference is to the large volume of sales on Singles Day.
- “Alibaba Cloud’s technical and operational expertise can certainly be applied in regions outside of China, but only until there’s customer demand and the data centers to serve it.”
Alibaba dumped American vendors as part of its journey.
Google Cloud
- “Google has arguably the only, truly global infrastructure, because its services and users are global.”
- “Google‘s services cannot anticipate traffic spikes, unlike a planned shopping holiday, and must be ready wherever, whenever it happens.”
- “Google’s products do not naturally lead to processing many complex transactions, like online shopping orders, offline delivery, or payments.”
Google can accommodate stress, but it’s not so good in Amazon-style transaction complexity.
Microsoft Azure
- “None of these [Microsoft] businesses have to be “always on”, in the same way that an e-commerce marketplace or a search engine needs to be on.”
- “Azure is still doing amazingly well from a revenue and market share standpoint. This success has more to do with Microsoft’s years of experience in selling products into large enterprises and aggressively moving users of its non-cloud license-based products onto the same products that are now on-cloud and subscription-based. Microsoft is very good at being “enterprise ready”, but not that good at being “Internet ready”.”
- “It [Microsoft] has by far the most number of Single-AZ Regions, which has led to outages and issues that could’ve been avoided with a multi-AZ design. Multi-AZ Region is the default in AWS, GCP, and most of Alibaba Cloud.”
Microsoft is good at sales, not so good at the cloud.
Net Net
Alibaba is darned good. At any time the company can push into other markets and create some pain for the American companies it seems.
Stephen E Arnold, March 10, 2020
Financial Companies: Cloud Innovators?
February 28, 2020
An online information service called Channel Life published results of a survey related to financial services (mostly in Australia) and the hybrid cloud. The phrase “hybrid cloud” means combining AWS or Azure like services with an organization’s own private cloud. The buzzword is a typical meta notion; that is, as features fragment, another method comes along to roll up these diverse approaches into a single umbrella service. Word and Excel became Microsoft Office which morphed into a subscription umbrella service. Now the same approach is finding favor in financial services.
The write up “Financial Organizations Leading the Way for Hybrid Cloud, Nutanix Finds” presents the results of a research “study.” Here are a handful of factoids DarkCyber finds interesting:
60% of respondents state security is the single biggest influence on future cloud strategies. If accurate, regulations drive adoption of technology, not a customer centric focus.
39% of financial services companies reporting public cloud services were completely meeting their expectations. If verifiable, commercial cloud services may be failing to meet some customers’ needs, a situation not evident from the marketing collateral from major cloud vendors.
18% of financial companies have deployed hybrid cloud, while 51% plan to shift investment to hybrid cloud in just three to five years. If verifiable, there’s money to be made in meta software.
DarkCyber spotted minimal information about the sample selection process, sample size, and statistical procedures used. Nevertheless, one message is that bright sprouts may be able to make money providing an integration layer for multiple cloud systems. Is the era of the branded cloud ending? No, but the cloud business may be ripe for evolution…. if the data in the report are generally correct.
Stephen E Arnold, February 28, 2020
Amazon Costs: Kubernetes to the Rescue
February 27, 2020
How much does an AWS customer pay for a specific service? DarkCyber knows that the taxi meter approach makes sense for the company that owns the medallions, the taxis, and possibly a few important people.
The taxi meter method in the cloud is a bit of a mystery — until the invoice arrives. “The Story Behind My Talk” explains a process which, according to Tuananh, can “reduce EC2 billing up to 80%.”
How does this money slip from the massive tractors of the Bezos bulldozer?
The article explains:
I started with a managed GKE at first using their free $300 credit, to learn the basic of Kubernetes; but then later on use kops to setup a production cluster with AWS.
Some of the changes I did for the production cluster is:
- Setup instance termination daemon to notify all the containers + graceful shutdown for all the apps.
- Setup multiple instance groups of various size and availability zone, mixing spot instances with reserved instances. This is to prevent price spike of certain spot instance group; and minimize the chances of all spot instances going down at the same time.
- Calculate and provision a slightly bigger fleet then what we actually need so that when there were instances shut off, there won’t be service downgrading. Because spot instances are so cheap, we can do this without worry much about the cost.
- Watch to see if there were scheduling failure to scale the reserved groups.
The payoff according to the article:
“The overall cost saving for EC2 was around 60-70% because we need to mix reserved instances in and provision a little higher than what we actually need. We were very happy with the result.”
Will AWS reconsider how it deals with EC2 billing? Does a bulldozer need diesel fuel?
Stephen E Arnold, February 27, 2020
Does Amazon Have Dark User Interface Patterns?
February 25, 2020
The question “Does Amazon make use of interfaces intentionally designed to generate revenue?” is an interesting one. Amazon does have a boatload of features, functions, and services. There are — what? — more than a half dozen different databases, including the quantum thing.
The article “My First AWS Free Tier Hosting Bill Was $900.” The idea is that “free” did not mean exactly free. This is akin to the word “unlimited” when it appears in mobile data plans. Is Amazon following a path blazed by telecommunications giants, truly models of consumer centric behavior in DarkCyber’s narrow view of the economic world.
The write up states:
A major part of AWS marketing is pay-per-use for their services:
“You only pay for the services you consume, and once you stop using them, there are no additional costs or termination fees.” – AWS Pricing
They also market “free tier” products, less powerful instances that are free for the first year of use.
The article reports that a slow roll out allowed the system to “sit around for a month.”
That decision cost about $1,000.
The article points out that assuming that an “idle server” would not cost anything. Also, the Amazon jargon did not make sense, so the developer ignored the Amazon speak.
The write up goes through the Amazon lingo to alert other individuals of Amazon’s approach to “free.”
Several observations:
- Amazon is confusing. DarkCyber thinks this is party due to the vaunted two pizza team approach to programming and part due to clever marketers who really want to match up to the founder’s principles.
- Amazon pitches itself hard as the logical, best, and superior choice for cloud anything. Individuals who buy this pig in a poke are going to pay.
- Amazon, if one makes a good case to the customer service unit staffed with people who sort of speak like those in rural Kentucky, will modify the charge.
Are these some lessons one can learn from this write up? Maybe, for example:
- Learn to speak Amazon
- Think before clicking
- Amazon became really big for a reason: Avoid becoming a third party merchant whose hot product became part of Amazon Basics.
Your mileage may vary from the drive through the Tunnel of Love that the author of the article took.
Stephen E Arnold, February 25, 2020
Protecting Digital Data Eggs with Multiple Clouds
February 20, 2020
Rumor has it the CIA has finally begun the procurement process to update its cloud technology, we learn from Inventa’s article, “The CIA Wants to Upgrade its Cloud Tech Without DoD’s JEDI Drama.” Though we’re told an agency spokesperson refused to confirm the plan, we imagine the CIA is eager to avoid the Iowa App syndrome. Writer surbhi suspects the CIA is reluctant to comment because it wishes to avoid the sort of spotlight that was cast last year on the Pentagon’s JEDI cloud procurement process. According to Nextgov, though, the agency has released its draft RFP, expects proposals to be in this spring, and plans to make a decision by September.
The article notes that, about a year ago, the CIA’s Directorate of Digital Innovation put forth its C2E (or Commercial Cloud Enterprise) plan. Though some specifics seem to have changed since that announcement, it likely still represents a commercial contract worth tens of billions of dollars. That impression is reinforced by the plan outlined last summer by the Director of National Intelligence, “The Strategic Plan to Advance Cloud Computing in the Intelligence Community” (PDF), which emphasizes:
“Information is exploding in volume and velocity and challenging our ability to expeditiously collect, analyze, and draw conclusions from disparate data sets. Additional manpower will not close the resulting gap; we must leverage leading edge technology. The future IC cloud environment presented herein will effectively function as a force multiplier to enhance our effectiveness and address mission challenges.”
The write-up reminds us:
“The CIA was an early adherent of the cloud when it chose Amazon to build a $600 million private cloud in 2013. … The Atlantic called it a ‘radical departure for the risk-averse intelligence community’ in a 2014 article. Cloud technology has certainly evolved in the seven years since the CIA last did this exercise, and it makes sense that it would want to update a system this old, which is really ancient history in technology terms. The CIA likely sees the same cloud value proposition as the private sector around flexibility, agility and resource elasticity, and wants the intelligence community to reap the same benefits of that approach. Certainly, it will help store, process and understand an ever-increasing amount of data, and put machine learning to bear on it as well.”
Putting the available digital eggs in one rucksack may not be prudent.
Cynthia Murrell, February 20, 2020
The Clouds in UAE: Amazon Not Mentioned
February 13, 2020
Here’s the big reveal in the write up titled “Microsoft Sees Room for Growth Opportunities for All Cloud Providers in UAE.”
The US technology giant [Microsoft] offers three main clouds – Azure, Microsoft Office 365 and Dynamics 365.
Three clouds. Well, Google has 10 chat apps. So much for efficiency, federation, and distributed architectures.
Other factoids in the write up, in DarkCyber’s opinion, are:
- The growth for Microsoft and “all” is PaaS or Platform as a Service.
- Competition is good, presumably among members of the oligopoly players in the “cloud”
- Alibaba has a cloud data center in the UAE
- The intelligent edge is a reality. What’s an intelligent edge? Hey, no need to explain this bit of verbal frippery.
- There are more than 1,000 technologies on Azure. Can anyone list these? DarkCyber cannot.
The write up does not mention the other outfit near Microsoft. But Amazon has some operations in the UAE, Bahrain, and other countries in the area as well.
The write up toots Microsoft’s tuba and Oracle’s flute.
Yep, useful marketing packaged as “real” analysis. But three “main clouds”?
Stephen E Arnold, February 13, 2020
Amazon: Some Trouble Down Under?
January 29, 2020
DarkCyber noted “Case Study: Why the Australian Electoral Commission Migrated to Microsoft Azure.” On the surface, the write up is another PR output. When considered in terms of the competition between Amazon and Microsoft for juicy non commercial jobs, the article provides a check list of what’s lacking in Amazon AWS. DarkCyber identified these “advantages” for the Redmond brain trust which finds questionable methods for altering a user’s Windows 7 machine amusing. (This black screen incident provides a reminder that PR check points may not match a firm’s actual behavior.)
Here are the upsides for Azure, presumably without a black screen on Luddites’ Windows 7 computers:
- Quick turn around
- Publicly exposed APIs
- API management tools
- API creation tools
- Real time information feeds
- Ability to create an “express route” for speedy data communications
- Zero failure
- Ability to support self service from users
- A customer or user service portal
- A much loved integrator.
What was the deciding factor? The much loved vendor it seems.
Does Amazon match up on these check points? Sure.
Marketing presentations are one thing. The much loved vendor is another.
Stephen E Arnold, January 29, 2020
Amazon Finds a Home in the UK
January 10, 2020
Just a quick item about Amazon Web Service contract size. “Home Office Reinforces Commitment to AWS with £100m Cloud Hosting Deal” makes clear that a UK government entity has not been won over my Microsoft Azure. The write up reports this information:
“The award of the public cloud hosting services contract to Amazon is a continuation of services already provided to the Home Office,” a departmental spokesperson told Computer Weekly. “The contract award provides significant savings for the department of a four-year term.” The Home Office is renowned for being a heavy user of cloud technologies, and is – according to the government’s own Digital Marketplace IT spending league table – by far the biggest buyer of off-premise services and technologies via the G-Cloud procurement framework.
The contract is significant because it suggests that other Five Eyes’ participants will be exposed to the AWS approach.
For Amazon staff working on the contract, there may be some meetings at Clarendon Terrance. London taxi drivers know where that is. No digital map needed.
Stephen E Arnold, January 10, 2020
Google: We Can Be Avis, National Car Rental or an Off Airport Outfit Too
December 18, 2019
Quite a goal. Google wants to beat Amazon and Microsoft in the cloud. Err, isn’t Google a cloud centric outfit, or at least since it morphed from the cutesy Backrub into the behemoth it is today? What if Google doesn’t think it is a cloud business? Hmm.
The answer, of course, is Googley. Google has waffled a bit. The phones, the home helpers, and the mouse pads. But the company operates “out there”, from data centers in regular buildings to wonky containers which can be towed to a location where power is cheap and skills are hard to come by.
A series of stories is zipping around about Google’s new desire to become the big dog in cloud computing. Just like the PR program featuring Jeff Dean, the Google is starting to realize that it may have more in common with the low rent business of scalping tickets than with high technology outfits changing the way business does business.
That’s an interesting thought because it runs counter to the received wisdom that Google is the font of technology. Like the fountains in Rome, lots of work is needed to keep the fountains spouting water. Tourists don’t see Rome’s plumbing, and for good reasons.
The goal of knocking off Amazon and / or Microsoft (love that lawyer conjunction, don’t you?) will be achieved by 2023. That works out to 24 months. Microsoft’s NT project turned into a death march, and I think this goal is likely to follow the same trajectory.
First, Amazon and Microsoft are not standing still. Good old Microsoft is working overtime to make Azure stable and semi-coherent. How many search engines does one desktop software company need? How many analytics solutions? How many servers? These are questions Microsoft engineers are rushing to answer. The airplane is aloft, and making adjustments to an engine when the plane in in flight can be difficult when it has to operate in a hybrid mode and the ground stations can be crashed by a software update. Cool?
Plus, Amazon is moving along a different trajectory. The company is engaged in a multi front war, and it is less and less a cloud company. That bookstore in Nashville and the undoing of FedEx make clear that not even a mid tier state like Tennessee is exempt from the Bezos bulldozer.
Second, Google has not been particularly adept at sticking with projects over time. Examples range from the social media attempts, to the Alon Halevy semantic tools, and to some as simple as messaging services. The culture of incompleteness is a hurdle. Managers can fiddle with incentives and tweak the hiring processes. But the company is a bit like a flotilla of sailboats generally heading toward port when a bad storm presents itself. Everyone knows where to go, but there may be some delays. Delays when trying to knock off Amazon and Microsoft may not be desirable.
Third, there are lots of other companies which want to be the Avis and National to the Uber business. Oracle, down but not out. IBM, a bit of a clueless geriatric but still capable of surprises like its sales success in India, and dozens upon dozens of other companies.
Net net: The write up “Google Brass Set 2023 as Deadline to Beat Amazon, Microsoft in Cloud” is useful, but it contains one telling statement:
Google shifted headcount growth to its cloud platform sales and engineering teams.
What’s going to be the Google equivalent of Windows 10 updates which don’t work, arrive late, and kill some data? If it is ad systems, Amazon is going to get the best location in the airport to serve rental car customers.
Stephen E Arnold, December 18, 2019
WWAD: What Will Amazon Do?
December 4, 2019
Silicon Angle published “Commentary: Andy Jassy Aims to Reinvent Amazon Web Services for the Cloud’s Next Generation.” The story carries the subtitle “In an exclusive one-on-one conversation, Amazon’s cloud chief reveals how he views the future of the cloud, the competition, market shifts, customer demands and controversies.”
Several statements in the write up warranted an orange highlight:
- It’s time to embrace the next cloud wave or get crushed by it.
- The cloud has completely “flipped the business and startup model on its head.”
- “Enterprises realize that if they want to be successful, sustainable companies over time, they can’t just make small, incremental changes,” he said.
- The “vast majority” of organizations pursuing a multicloud strategy tend to pick a predominant provider and then, if they feel like they want another one, either because there’s a group that really is passionate about them or they want to know they can use a second cloud provider in case they fall out of sorts with the initial cloud provider, they will. Jassy went on to say that for customers implementing multiple clouds the workloads are split between a primary and secondary cloud more like 70/30 or 80/20 or 90/10, not 50/50.
- “Companies are going to want to eliminate network hops and find a way to have the compute and the storage much more local to the 5G network edge.”
- Next year roughly 82% of all new workloads will run Linux.
Net net: Crushing is part of the game plan. The interview is a component of the AWS re:Invent PR push. Prime stuff, not Grade A, but okay for consumption by Amazon shoppers.
Stephen E Arnold, December 4, 2019
-
- Subscribe to Beyond Search
Feature archive
News archive
-
