If you control your code base and you have only a handful of applications that run at massive scale – what some have called hyperscale – then you, too, can win the Chip Jackpot like Meta Platforms and a few dozen companies and governments in the world have. …
With MTIA v2 Chip, Meta Can Do AI Inference, But Not Training was written by Timothy Prickett Morgan at The Next Platform.
The steady rise of AI over the past several years – and the accelerated growth with the introduction generative AI since OpenAI’s launch of ChatGPT in November 2022 – has shifted Intel’s status as a challenger in a chip market that it long had dominated. …
Gelsinger: With Gaudi 3 and Xeon 6, AI Workloads Will Come Our Way was written by Jeffrey Burt at The Next Platform.
At this stage of the build internally there is only really the bedroom, snags and trims (windows, ceiling edges and doors) left to do. I am currently in Australia having a bit of RnR so this is a reflective post to show the build at its present state and go through the things I have learnt along the way. I could plan all I like but as I haven’t lived on a boat before there was always going to be a few wrong design decisions.
Dmitry Perets wrote an excellent description of how typical firewall cluster solutions implement control-plane high availability, in particular, the routing protocol Graceful Restart feature (slightly edited):
Most of the HA clustering solutions for stateful firewalls that I know implement a single-brain model, where the entire cluster is seen by the outside network as a single node. The node that is currently primary runs the control plane (hence, I call it single-brain). Sessions and the forwarding plane are synchronized between the nodes.
In the previous three parts, we learned about all the interesting things that go on in the PHY with scrambling, descrambling, synchronization, auto negotiation, FEC encoding, and so on. This is all essential knowledge that we need to have to understand how the PHY can detect that a link has gone down, or is performing so badly that it doesn’t make sense to keep the link up.
The function in 1000BASE-T that is responsible for monitoring the status of the link is called link monitor and is defined in 40.4.2.5. The standard does not define much on what goes on in link monitor, though. Below is an excerpt from the standard:
Link Monitor determines the status of the underlying receive channel and communicates it via the variable
link_status. Failure of the underlying receive channel typically causes the PMA’s clients to suspend normal
operation.
The Link Monitor function shall comply with the state diagram of Figure 40–17.
The state diagram (redrawn by me) is shown below:
While 1000BASE-T leaves what the PHY monitors in link monitor to the implementer, there are still some interesting variables and timers that you should be Continue reading
If you are wondering why Intel chief executive officer Pat Gelsinger has been working so hard to get the company’s foundry business not only back on track but utterly transformed into a merchant foundry that, by 2030 or so can take away some business from archrival Taiwan Semiconductor Manufacturing Co, the reason is simple. …
Google Joins The Homegrown Arm Server CPU Club was written by Timothy Prickett Morgan at The Next Platform.
Data exfiltration and ransomware attacks in cloud-native applications are evolving cyber threats that pose significant risks to organizations, leading to substantial financial losses, reputational damage, and operational disruptions. As Kubernetes adoption grows for running containerized applications, it becomes imperative to address the unique security challenges it presents. This article explores the economic impact of data exfiltration and ransomware attacks, their modus operandi in Kubernetes environments, and effective strategies to secure egress traffic. We will delve into the implementation of DNS policies and networksets, their role in simplifying egress control enforcement, and the importance of monitoring and alerting for suspicious egress activity. By adopting these measures, organizations can strengthen their containerized application’s security posture running in Kubernetes and mitigate the risks associated with these prevalent cyber threats.
Data exfiltration and ransomware attacks have emerged as formidable threats to organizations worldwide, causing substantial financial losses and service outage. According to IBM’s 2023 Cost of a Data Breach report, data exfiltration attacks alone cost businesses an average of $3.86 million per incident, a staggering figure that underscores the severity of this issue. Ransomware attacks, on the other hand, can inflict even more damage, with Continue reading
We have said it before, and we will say it again right here: If you can make a matrix math engine that runs the PyTorch framework and the Llama large language model, both of which are open source and both of which come out of Meta Platforms and both of which will be widely adopted by enterprises, then you can sell that matrix math engine. …
With Gaudi 3, Intel Can Sell AI Accelerators To The PyTorch Masses was written by Timothy Prickett Morgan at The Next Platform.
A long-time friend sent me this question:
I would like your advice or a reference to a security framework I must consider when building a green field backbone in SR/MPLS.
Before going into the details, keep in mind that the core SR/MPLS functionality is not much different than the traditional MPLS:
There are events that unite people, like a total solar eclipse, reminding us, humans living on planet Earth, of our shared dependence on the sun. Excitement was obvious in Mexico, several US states, and Canada during the total solar eclipse that occurred on April 8, 2024. Dubbed the Great North American Eclipse, millions gathered outdoors to witness the Moon pass between Earth and the Sun, casting darkness over fortunate states. Amidst the typical gesture of putting the eclipse glasses on and taking them off, depending on if people were looking at the sky during the total eclipse, or before or after, what happened to Internet traffic?
Cloudflare’s data shows a clear impact on Internet traffic from Mexico to Canada, following the path of totality. The eclipse occurred between 15:42 UTC and 20:52 UTC, moving from south to north, as seen in this NASA image of the path and percentage of darkness of the eclipse.
Looking at the United States in aggregate terms, bytes delivered traffic dropped by 8%, and request traffic by 12% as compared to the previous week at 19:00 UTC Continue reading
We have been tracking the financial results for the big players in the datacenter that are public companies for three and a half decades, but starting last year we started dicing and slicing the numbers for the largest IT suppliers for stuff that goes into datacenters so we can give you a better sense what is and what is not happening out there. …
Mixed Results For The Datacenter Thundering Thirteen In Q4 was written by Timothy Prickett Morgan at The Next Platform.