The open source, AI-native proxy, created by @soloio_inc to optimize connectivity, security and observability for AI agentic software, has joined the @linuxfoundation.
By @sjvn https://t.co/UfawpLRexz
The open source, AI-native proxy, created by @soloio_inc to optimize connectivity, security and observability for AI agentic software, has joined the @linuxfoundation.
By @sjvn https://t.co/fA1sJ5wUNh
@serathius@Itsuugo Oh but I guess you don't want the full encoded thing in memory at once. So you probably would need some sort of lazy Buffer implementation maybe - not sure it would work end to end
@serathius@Itsuugo Basically you can give it [][]byte instead of []byte when you encode now (https://t.co/o4N9Is4by1). In my case we have a repeated `Any` so we not only wanted to split up the bytes but not have to copy them (Any already is basically just wrapping []byte)
@xu_paco@wzshiming Nice! Curious, how is it beta if the containerd support just merged? Was it implemented in a different way before and now can use native containerd implementation or something?
Today marks exactly 6 years from my first commit of a long journey to make @IstioMesh scale.
When I first started, a 100 pod mesh was a "big" mesh. Today, we can handle anything - 100 *million* pods and beyond!
Service mesh scalability is a solved problem.
Link 👇
Top KubeCon question: "Does ambient mesh scale?"
Yes, but a more useful question is "How much effort does it take to scale"?
https://t.co/brv2OAhT9D covers this in depth.
@cmanolache@linsun_unc Yep exactly, goal was to show HTTP processing overhead (or lack of overhead, really).
The aggregate data throughput here is not huge, something like 50Mb/s or so -- bottleneck here is more latency than throughput since its a bunch of tiny messages.
Powered by Gloo Ambient mesh, I was able to push through *410k requests per second*, with mTLS encryption and HTTP monitoring.
All served by a single 8-core CPU (on GCP).
https://t.co/4OvTYblDR0 Ambient mode — the next-generation architecture for Istio — more than just an incremental improvement to the sidecar-based architecture. #ambientmode#istio#kubecon#multicloud