Microsoft, AT&T Sign Cloud Deal Worth More Than $2 Billion (reuters.com) 26
Microsoft and AT&T on Wednesday said they reached a deal under which the telecommunications company will tap Microsoft's Azure cloud service for its computing needs and use Microsoft 365, which includes Office productivity software, for much of its 268,000-strong workforce. From a report: Under the deal, Microsoft and AT&T will also work together on so-called edge computing, which will see Microsoft technology deployed alongside AT&T's coming 5G network for applications that need extremely small delays in passing data back and forth, such as air traffic control systems for drones. The multi-year deal is worth more than $2 billion, according to a person familiar with the matter. The agreement is a major win for Microsoft, which will become AT&T's "preferred" cloud vendor and is fighting to gain market share from Amazon Web Services, the biggest provider of public cloud services. Cloud service customers run their software applications in data centers managed by the cloud provider. AT&T will remain responsible for its own core networking operations for cell phones and other devices. But John Donovan, chief executive of AT&T Communications, told Reuters the deal is a fundamental shift for the telecommunications provider to become "public cloud first," meaning that it will predominately rely on data centers built by others to power the rest of its business.
The Empire contracts with Darth Vador (Score:2)
Good luck with the death star deployment guys...
Re: (Score:2)
You... You.... You misspelled Darth Vader... On a site for nerds... I've never felt so sad. ;_;
"Cloud" (Score:1)
Re: (Score:3)
It's funny, because I see vendors already pushing backup products for MS cloud products because the "backups" they do don't cover your own loss of data due to all the usual fuckups which cause data loss.
Plus there are no "serverless" apps, they all run on servers *somewhere*, you're just paying a premium for someone else to do the racking, DBA, networking, etc. And unless your app is exclusively developed and deployed via the cellular network, somewhere there's a network, and possibly an elaborate one, dep
Re: (Score:2)
>Plus there are no "serverless" apps, they all run on servers *somewhere*, you're just paying a premium for someone else to do the racking, DBA, networking, etc
Semantics, terminology overloading.
First of all, what is a "server"? Is there any agreement on what it means? Is it a computer in a box that goes in a rack? Is it a VM? Or is it a software application that communicates with clients? The term is ambiguous.
"serverless" is perhaps an unfortunate terms, but I would say that FAAS (functions as a servic
Re: (Score:2)
A big chunk of this just seems to be ridiculous devolution, depending on how invested you are in whatever the latest going trend is for microsegmentation.
Re:"Cloud" (Score:5, Interesting)
Oddly enough I found that lower-middle management loves using these terms to impress the MBA's more then what the MBA's like to use.
That being said, I Remember when "Cloud" Computing was something different then it is today. It use to be used for a Distributed computing methodology (Think SETI At home) where People PC, which are often well overpowered for most uses, would be hooked in the network, and aid in processing.
While the principals have been roughly applied, it is far more centralized in a data center. However what most people call "Cloud" use to be Called "SaaS" which used to be called "Hosted Services" which use to be called "Time Sharing Computing" where they all fall under the following model. Powerful Computers are expensive, and require a lot of work to keep functional. So lets just pay someone a monthly fee and they cover the upfront expense of buying all the expensive computers.
Back in the Time Sharing Days, Mainframes and even Mini-computers were extremely expensive for most small - mid sized businesses, and what needed to be calculated was only needed for a fraction of the system possible performance.
strong workforce (Score:1)
268,000 strong workforce. not anymore.
How the mighty have fallen - from reliability. (Score:3)
Back around the turn of the millenium I was working on Unix disk-mirroring drivers for the company that sold mainframes into all the telecoms. They were really concerned about uptime (as well as quick recovery), because if these machines went down all the toll calls on their network were free until they came back up. The millions of lost dollars add up amazingly fast.
We had one patch (in another driver) that had been in an auto-install queue for YEARS waiting for the machine to reboot. It had a too-small counter and you had to hot-poke a process when it rolled over. When I was there it had rolled over TWICE and we were still waiting....
I suppose the billing isn't as big an issue now that so much of telephony and other networking is flat-rate. But other aspects of its networking still are: Telephony is critical to emergency services. So if something fails that keeps calls from going through, people DIE.
Because of this, the specs on their hardware was tougher than Mil. And it carried over into their software, and other aspects of their operation, too. It was one of the things that fed into the "Bell head / Net head" conflict and resistance to IP-style packet switching: Telephone networks were all about "No bit shall fall", and when they did get into packet-switched designs they involved reserved bandwidth, per-hop error handling, and tight timing control. The idea of building a network where packets of bits from a flow take individual routes, and often just get lost and have to be retransmitted if they really mattered, literally made their heads hurt.
So now they're outsourcing some of their computing to Microsoft? Good grief! I know that there have been a few changes of management since those days. But I hope there's still enough of their historic concern for reliability left that they're careful to keep the life-critical stuff in-house, where they can keep its reliability under their own control - and close scrutiny.
IBM + Microsoft = less Amazon (Score:2)
When I saw this, I thought I was seeing double! Because the day before this announcement, ATT and IBM announced a similar agreement - worded almost identically, save for "preferred" cloud vendor (Microsoft) and Office 365. There was some language in the IBM announcement about moving workloads from inhouse to cloud. The numbers are vague, but it SEEMS the IBM contract may be the larger one.
My GUESS is that ATT will start moving some workloads from AWS to Azure, while at the same time moving some workloads fr
Re: (Score:2)
The timing of the IBM/ATT announcement corresponds pretty closely to IBM's announcement of Hyper Protect DBaaS for PostgreSQL and MongoDB.
This is running PostgreSQL/MongoDB clusters on highly-secure IBM Z-series (LinuxOne) servers. Something that only IBM can offer vertically, and to my knowledge is currently not available on either AWS or Azure. If if they offered it, IBM would derive significant hardware revenue. If you have data with a high security requirement, do you want you database hosted on Intel h