Amazon appears to be planning to further expand its datacenter footprint in the US to meet growing demand for cloud computing services.
Based on a new proposal (PDF) submitted to the City of Gilroy in Santa Clara County, California, the cloud computing giant is planning to build a new site that will eventually house two datacenter buildings.
These new data centers would expand AWS availability on the West Coast and each is planned to have a capacity of up to 49MW of equipment plus a security building and backup generator facilities according to a new report from The Register.
Amazon’s new Gilroy Data Center (GDC) facility in California will be constructed in two phases if given the go-ahead by the City of Gilroy.
The company expects the first phase which will consist of a single-story datacenter building, an onsite substation, a security building and a battery energy storage system to be operational by 2023. Phase two will involve creating another single-story datacenter building that is 218,000 square feet and it is expected to be built within four to seven years following the completion of phase one depending on customer demand.
While the phase one data center will use diesel-fired backup generators, Amazon intends to use alternative backup generation technologies for its phase two data center. One of the reasons for this is due to the fact that the company’s new facility will be located on farmland just outside of Gilroy that it purchased back in 2020 for $31.3m.
Based on local media reports though, the City of Gilroy is in the process of soliciting bids from outside consultants in an effort to determine the potential environmental impact of Amazon’s Gilroy Data Center facility before granting approval for the project.
We’ll likely hear more from Amazon if and when its new Gilroy Data Center facility is approved and moves on to the construction phase.
A potentially major security flaw has been discovered on Rarible, a popular marketplace for non-fungible tokens (NFT), which could lead to users losing not just their NFTs, but also the cryptocurrencies right from their wallets.
A report from Check Point Research (CPR) identified a vulnerability that would allow a potential attacker to steal someone’s digital belongings in a single transaction. The worst part is that everything would happen on the marketplace itself, a place people would generally feel less suspicious.
According to CPRs report, the methodology is simple, and includes creating a “malicious NFT”. Should someone stumble upon it, and click on it, the malicious NFT would execute JavaScript code in an attempt to send a setApprovalForAll request to the victim.
In case the victim submits the requests, they’d grant the malicious NFT full access to their endpoint.
“In October last year, we discovered critical security flaws in OpenSea, the world's largest NFT marketplace. Now, we've identified similar vulnerabilities in Rarible,” commented Oded Vanunu, Head of Products Vulnerabilities Research at Check Point Software.
“In terms of security, there is still a huge gap between Web2 and Web3 infrastructure. Any small vulnerability opens a backdoor for cybercriminals to hijack crypto wallets behind the scenes. We are still in a state where marketplaces that combine Web3 protocols are lacking a sound security practice. The implications following a crypto hack can be extreme. We've seen millions of dollars hijacked from users of marketplaces that combine blockchain technologies.”
Last year, Rarible has had more than $273 million in trading volume, making it one of the largest NFT marketplaces on the planet.
The company notified the marketplace of its discovery, and said it “believes Rarible will have deployed a fix by the time of this publication”. We have reached out to Rarible to see if that indeed is the case, and will update the article accordingly.
However, given that it’s Easter weekend, it could be a few days before we hear back from Rarible.
“Users currently need to manage two types of wallets: one for most of their crypto and another just for specific transactions,” Vanunu continued.
“Should the wallet for specific transactions become compromised, users can still be in a position where they don’t lose everything."
Following the release of version 100 of its browser, Google has released a new update for Chrome to fix a high-severity zero-day vulnerability that is being actively exploited in the wild.
According to a new security advisory put out by the search giant, the company is aware that an exploit exists for this high-severity vulnerability tracked as CVE-2022-1364.
The bug itself is a confusion weakness in the Chrome V8 JavaScript engine and while these types of vulnerabilities usually lead to browser crashes after reading or writing memory out of buffer bounds, cybercriminals can also exploit them to execute arbitrary code on vulnerable systems.
The vulnerability was discovered by Clément Lecigne from Google’s Threat Analysis Group who immediately reported it to the Google Chrome team. Although Google has observed this zero-day actively being exploited in the wild, the company has been tight lipped regarding any attacks. In its security advisory, it said that details about the bug and links will be “kept restricted until a majority of users are updated with a fix”.
(Image credit: Google)
Manually updating Chrome is your bet bet
Google Chrome 100.0.4896.127 for Windows, Mac and Linux will roll out in the next few weeks as an update.
However, due to the high-severity of this vulnerability, security-conscious users can update Chrome immediately by going into the Chrome menu, heading to Help and clicking on About Google Chrome. Here, they’ll be able to manually install the update themselves as opposed to waiting for Google to roll it out.
For those that would rather wait though, Chrome will automatically check for new updates and install them the next time you close and relaunch the browser.
This is the third zero-day vulnerability that has been discovered and patched in Chrome this year.
A revamped Apple Mac Mini for 2022 is hotly anticipated, so like you, TechRadar's editors were surprised it didn't appear at Apple's Peek Performance event on March 8.
We did see something else for fans of the Apple desktop line: the Mac Studio, featuring the new M1 Ultra SoC (system-on-a-chip), designed with creative professionals in mind. But something is clearly brewing: Convincing new rumors say Apple is testing nine new Mac models ahead of WWDC with brand-spanking new M2 silicon chips -- and that includes a new Mac Mini.
The mini-PC impressed us when we reviewed it in 2020, and the M1 chip powering it showed itself to be a pretty competent performer overall. Still, a home PC is a pretty big piece of kit. Can a revamped 2022 Mac Mini win over converts beyond the device's fanbase? Here's everything we know so far.
Apple Mac Mini (2022): cut to the chase
What is it? Apple's next-generation mini-PC
When is it out? Expected late 2022 or 2023
What will it cost? Unknown, but likely similar to current pricing (starting at $699 / £699 / AU$1,099)
April 14: A firmware update for Apple’s new Studio Display references a product marked as Macmini10,1, which apparently is a model ID not used for any of the existing Mac product lines.
March 13: Bo-ring! Well-respected and (usually very well-informed) Apple analyst Ming-Chi Kuo says the upcoming Mac Mini will keep the existing form factor - and will show up in 2023. :-(
Feb. 21: Seven new Mac models will be released in 2022, says Mark Gurman, all featuring some version of the Apple silicon processors. Including a new Mac Mini.
March 04: With an Apple event right around the corner, there's a lot of attention being paid to rumors of new MacBooks, iPads, and iPhones. But a new M2-powered Mac Mini could be the star in the show, writes TechRadar's John Loeffler.
Aug. 23, 2021: Bloomberg’s Mark Gurman suggests that the M1X Mac mini could launch “in the next few months.” Could that mean Apple will have a massive announcement event planned for September?
Release date and price
(Image credit: Future)
The Apple Mac Mini (2022) hasn't even been announced yet, so there's no official release date that we can confirm. The last Mac Mini (M1) was released in late 2020, so going off Apple's typical 18 to 24 product release cadence, we expected it to be announced at the March 2022 event, but sadly this wasn't the case. It's still likely we could see it announced before the end of the year though if Apple does have plans to release an M2 SoC.
The 2020 models of both products were announced at the same time, and both feature the M1 chip. There are also a lot of rumors that the 13-inch MacBook Pro won't see a major redesign, which might be reserved for a new MacBook Air reveal later this year.
If that is true, then it may make sense for the Mac Mini with M2 to be announced alongside the 13-inch MacBook Pro since there's much less excitement around a Mac Mini redesign than there is for a MacBook.
After all, announcing two new M2-powered devices that don't feature major redesigns is better than announcing just the one, and if the Mini were to only get internal improvements, swapping out the M1 for the M2 would make mass production significantly simpler.
More recent rumors from well-placed sources suggest that Apple might have decided to push back the Mac Mini 2022 to 2023 – and if that turns out to be the case, we're going to have to stop referring to it as the Mac Mini 2022.
Design
(Image credit: Future)
One of the major advantages to the Mac Mini is its size, which is perfect for those who need a home PC but don't want a large PC case or an all-in-one computer like the latest iMac. But will that change?
There has been a lot of talk about the potential redesign of Apple's MacBook Air (2022) as well as the lack of redesign for the 13-inch MacBook Pro. Which side of things do we expect the Mac Mini to come down on?
It's a hard question to answer, unfortunately: since there is less interest in seeing a Mac Mini redesign, there's less incentive for Apple to redesign the Mac Mini. However, one of the reasons why the lack of a redesign for the 13-inch MacBook Pro makes a lot of sense is that the 13-inch MacBook Pro might be discontinued after this year in favor of the 14-inch MacBook Pro (2021).
If you're going to continue to release a product, and we don't expect the Mac Mini to be discontinued any time soon, then it's almost certain that it will have to get a redesign (albeit, nothing major) to fit in with other Mac products that do, namely the iMac, and presumably the MacBook Air and MacBook Pro 14-inch and MacBook Pro 16-inch.
The rumor mill has gone both ways on this, but earlier leaks hinting that a change in form factor was on the cards have since been shot down by other sources. We'll have to wait and see what the Mac Mini 2022 looks like.
Specs and performance
There's not much we can say specifically about the Mac Mini's specs without knowing the Apple chip that runs it, but we do expect that it will have the same 8GB/16GB Unified Memory configuration options, with anywhere from a 256GB SSD up to a 2TB SSD.
We also expect the number and types of ports to stay more or less the same, but we're hoping for an upgrade to USB-C Thunderbolt 4 from the existing Thunderbolt 3. Same goes for upgrades in the way of HDMI 2.1 and a DisplayPort 1.4 output, but with Thunderbolt 4, those ports might not be necessary anyway, so they might get the axe with a new Mac Mini.
That might shift around our list of best monitors for the Mac Mini, but something tells us this isn't much of a concern for Apple.
As for the chips powering the new device, that's where things can get interesting.
There is some buzz that we won't just see an M2-powered Mac Mini, but we might also see an M1 Pro and even an outside chance of an M1 Max-powered Mac Mini.
These latter two might be more for professionals rather than home PCs, but the possibility could give the Mac Mini much broader appeal, especially since it might allow consumers of all stripes to access the performance of an M1 Pro and M1 Max at a more accessible price point than you'd get with a MacBook Pro.
Cerebras Systems, makers of the world’s largest chip, has announced that its CS-2 system now supports PyTorch and TensorFlow which will make it possible for researchers to quickly and easily train models with billions of parameters.
The company’s CS-2 is the world’s fastest AI system and is powered by its Wafer-Scale Engine 2 (WSE-2) CPU. With the release of version 1.2 of the Cerebras Software Platform (CSoft), the CS-2 now supports additional machine learning frameworks which will give developers even more choice when it comes to the types of models they want to run.
Senior director of AI framework at Cerebras Systems, Emad Barsoum provided further insight in a press release on how CSoft now enables developers to express models written in either TensorFlow or PyTorch, saying:
“From the start, our goal was to seamlessly support whichever machine learning framework our customers wanted to write in. Our customers write in TensorFlow and in PyTorch, and our software stack, CSoft, makes it quick and easy to express your models in the framework of your choice. By doing so, our customers gain access to the 850,000 AI optimized cores and 40 Gigabytes of on-chip memory in the Cerebras CS-2.”
Scaling large language models
CSoft version 1.2 now enables developers to write their models in the open source frameworks of PyTorch or TensorFlow and run them on the Cerebras CS-2 without any modification whatsoever. At the same time, an AI model written for a GPU or CPU can run in CSoft on the CS-2 without any changes.
With the combined power of CS-2 and CSoft, developers can seamlessly scale up from small models such as BERT to the largest models in existence like GPT-3.
Training large models using a GPU is challenging and time-consuming while training from scratch on new datasets often takes weeks and 10s of megawatts of power on large clusters of legacy equipment. Additionally, as the size of the cluster grows, power, cost and complexity grow exponentially.
Cerebras Systems built the CS-2 to address these challenges and its AI system can set up even the largest models in only a few minutes. Since developers spend less time setting up, configuring and training their models with the CS-2, they are able to explore more ideas in even less time.
