With new generations of GPUs and other kinds of AI accelerators either shipping or soon to start shipping and new CPUs also soon to be available from Intel and AMD, and sales already at a historical high level at Supermicro, you might not be expecting for sales to bust through a whole new higher ceiling starting in the next quarter. …
Supermicro Finally Mints Some Coin Peddling Rackscale Iron was written by Timothy Prickett Morgan at The Next Platform.
When I decided to sunset the ipSpace.net subscription, I asked Rachel Traylor whether I could make the content of her webinars public. She graciously agreed, and the first results are already here: you can download approximately half of the videos from the Network Connectivity and Graph Theory without a valid ipSpace.net account. Enjoy!
It is not a coincidence that the companies that got the most “Hopper” H100 allocations from Nvidia in 2023 were also the hyperscalers and cloud builders, who in many cases wear both hats and who are as interested in renting out their GPU capacity for others to build AI models as they are in innovating in the development of large language models. …
How To Make More Money Renting A GPU Than Nvidia Makes Selling It was written by Timothy Prickett Morgan at The Next Platform.
Is Open Source Software (OSS) a market failure? What does OSS add to the market that cannot be accomplished in other ways? What happened to the F (Free)? Join us for this roundtable episode of the Hedge.
Instead of being a protocol, EVPN is a solution that utilizes the Multi-Protocol Border Gateway Protocol (MP-BGP) for its control plane in an overlay network. Besides, EVPN employs Virtual eXtensible Local Area Network (VXLAN) encapsulation for the data plane of the overlay network.
Multi-Protocol BGP (MP-BGP) is an extension of BGP-4 that allows BGP speakers to encode Network Layer Reachability Information (NLRI) of various address types, including IPv4/6, VPNv4, and MAC addresses, into BGP Update messages. The MP_REACH_NLRI path attribute (PA) carried within MP-BGP update messages includes Address Family Identifier (AFI) and Subsequent Address Family Identifier (SAFI) attributes. The combination of AFI and SAFI determines the semantics of the carried Network Layer Reachability Information (NLRI). For example, AFI-25 (L2VPN) with SAFI-70 (EVPN) defines an MP-BGP-based L2VPN solution, which extends a broadcast domain in a multipoint manner over a routed IPv4 infrastructure using an Ethernet VPN (EVPN) solution.
BGP EVPN Route Types (BGP RT) carried in BGP update messages describe the advertised EVPN NLRIs (Network Layer Reachability Information) type. Besides publishing IP Prefix information with IP Prefix Route (EVPN RT 5), BGP EVPN uses MAC Advertisement Route (EVPN RT 2) Continue reading
If you insist on building your network with EBGP as a better IGP, make sure your implementation supports running IPv4 and IPv6 address families over EBGP sessions established between IPv6 link-local addresses (the functionality lovingly called unnumbered EBGP sessions).
Want to practice that neat trick? Check out the EBGP Sessions over IPv6 LLA Interfaces lab exercise.
If you insist on building your network with EBGP as a better IGP, make sure your implementation supports running IPv4 and IPv6 address families over EBGP sessions established between IPv6 link-local addresses (the functionality lovingly called unnumbered EBGP sessions).
Want to practice that neat trick? Check out the EBGP Sessions over IPv6 LLA Interfaces lab exercise.
Observability, especially in the context of cloud-native applications, is important for several reasons. First and foremost is security. By design, cloud-native applications rely on multiple, dynamic, distributed, and highly ephemeral components or microservices, with each microservice operating and scaling independently to deliver the application functionality. In this type of microservices-based architecture, observability and metrics provide security insights that enable teams to identify and mitigate zero-day threats through the detection of anomalies in microservices metrics, such as traffic flow, process calls, syscalls, and more. Using machine learning (ML) and heuristic analysis, security teams can identify abnormal behavior and issue alerts.
Observability also enables security teams to visualize the blast radius in the event of a breach. Using this information, teams can apply mitigating controls, such as security policy updates, to isolate the breached microservice and thereby limit exposure.
And finally, observability helps DevOps teams maintain the quality of service by identifying service failure and performance hotspots, and conducting a detailed investigation with capabilities such as packet capture and distributed tracing.
DevOps and SRE teams today are being overwhelmed by an enormous amount of data from multiple, disparate systems that monitor infrastructure and Continue reading
A few years ago, it was hard to imagine how AMD would have survived without re-entering the datacenter with its CPU and GPU compute engines. …
AMD Firing On All Compute Engine Cylinders was written by Timothy Prickett Morgan at The Next Platform.
In my Python journey, I've always stuck to just one version of Python at a time. I happily used Python 3.9 for quite a while, then switched to Python 3.10 without any issues. Everything was perfect until recently when I tried installing a Python application using pip, but no matter what I did, the installation kept failing. I couldn't fix the issue even after hours of Googling.
That's when I finally decided to check the documentation (which I should have done from the start), and there it was, this application requires a minimum Python version of 3.8 and was only tested on versions 3.8 and 3.9. That made me think, maybe I should have installed it using Python 3.9, but how? I'm no expert in Linux or Unix systems, and I worried that reinstalling Python 3.9 could mess up other projects I'd built on 3.10.
So, I started exploring how to manage multiple Python versions on the same machine, and that's when I stumbled upon a tool called 'pyenv'. This seemed like the perfect solution to my problem, so I decided to learn more about it.
When building leaf and spine networks, leafs connect to spines, but leafs don’t connect to leafs, and spines don’t connect to spines. There are exceptions to this and vPC is one of those exceptions. The leafs that are going to be part of the same vPC need to connect to each other. There are two ways of achieving this:
We will first use physical interfaces and then later remove that and use fabric peering. Now, my lab is virtual so take physical with a grain of salt, but they will be dedicated interfaces. The following will be required to configure vPC:
This is shown below:
The vPC is now up:
Leaf1# show vpc Legend: (*) - local vPC is down, forwarding via vPC peer-link vPC domain id : 1 Peer status : peer adjacency formed ok vPC keep-alive status : peer is alive Configuration consistency status : success Per-vlan consistency status : success Type-2 consistency status : success Continue reading
Pi-hole, the network-wide ad-blocking solution, has recently undergone significant updates, introducing a range of new […]
The post Testing Pi-hole Development Version 6 first appeared on Brezular's Blog.
When it comes to solving data analytics problems at scale, it is tough to beat the hyperscalers. …
The Perfect AI Storage: Trino From Facebook And Iceberg From Netflix? was written by Timothy Prickett Morgan at The Next Platform.