enhances Workers w stateful objects. This opens a huge new use cases for app dev. blog.cloudflare.com/introducing-wo…. Blog post from @KentonVarda does a good job in showing its potential in laymen’s terms. This is pretty exciting from a developer stand point. /1
This is way more impactful than the “edge db” I was guessing (a stateful data store). This is stateful OBJECTS — instead of having a shared db across Worker instances, the Worker instance itself is what is shared. /2
Basically, shared objects can be maintaining state as a mini in-mem storage and logic pod. Think of it as a shared spreadsheet or doc. With this new feature, each doc is a separate Worker instance - you aren’t storing the data in a shared (multi-tenant) database or API layer. /3
This is flipping “serverless” on its head a bit. Serverless & microservices are about enabling and scaling “stateless” requests (each request is its own thing, nothing is remembered from request to request). /4
“Stateful” = the object remembers (persists) its data between requests. You can now have multiple requests all sharing and acting on the same pool of logic and data. This is the “old server-client” method but brought into the next century. /5
Add in the scale and portability and geolocationality you get from the edge. Services can spin up countless objects when and where they need them. (As usage demands.) /6
This is pretty cutting edge stuff from a development angle. But easy to see potential in any pockets of intercommunication between any number of users. /7
1: Video sharing/broadcasting where each shared feed is a Worker. Zoom (or edge-based competitor!) could spin up endless serverless instances to each serve its own feed, geolocated across edge. /8
2: Communal gaming or gambling where each match is its own Worker. Matches could be spun up and down as played, running nearest edge to the globally situated users. /9
3: Collaborative editing. Shared objects as separate Worker instances could be being edited in real-time in its own edge compute space, apart from all others. /10
Those are just a glimpse at the potential. Any Workers dev could spin up billons of shared stateful objects across the globe. This greatly changes how software developers approach edge compute. /11
Excited to see this play at from here, esp paired w the power of programmable Edge Networks. Edge compute is now its own thing w very diff feature set than core cloud infra, to enable some incredible next-gen capabilities.
For now, it is in private beta. /fin
P.S. I am gonna count this as a correct prediction yesterday of “edge database”... but it is all much more potent and exciting then I imagined. This is the future of global software development.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Another quarter, another chance for Palo Alto $PANW to completely change up how it tracks customer counts within its Next-Gen Security segments.
This company cannot decide on what the hell to report for custs on Prisma SASE (Zero Trust), Prisma Cloud (CNAPP), and Cortex (XDR).
This time in Q4, they morphed from "pure" SASE customers (Access + Cloudgenix) to now include PANOS SD-WAN (their other SD-WAN before Cloudgenix acquire) and "SaaS Security subs" (WTF that is).
So they last revised this KPI in Q1, and now revise it AGAIN in Q4. WTF!
All this does it further bury their Access (Zero Trust) customers in a sea of SD-WAN custs and now broader platform SaaS subs.
And if you want to compare the 3556 SASE customers reported in Q4, look to the footnotes in the Q3 and Q2 slides (3160 and 2915). (Sorry, no Q1!)
The lack of CEO experience worked out so well last time they decided to go for it again. Let's look at his background...
He came out of their Meraki division, which Cisco acquired back in 2012 for on-prem Wifi & network appliances that are managed centrally via a cloud-based platform. It has expanded into SD-WAN appliances, connected IoT, and mobile device mgmt. ...
Of course, an edge network is one giant software-defined network. But not sure how this appliance-heavy experience is relevant to running a global edge network with CDN, WAF, & serverless computing capabilities. ...
... and hints of a few small things coming this week
"... driving down costs of core cloud services, pushing the boundary of our network to our customers' doorsteps, and investing in new technologies that may someday disrupt the web as we know it today"
which to me, reads as...
- transferring data between clouds & across the edge
- spreading the reach of the edge, like the Borg
- embracing web3/decentralization (as just stated on CloudflareTV kickoff by @eastdakota)
An investment strategy: buy and forget
Another: buy and shift (from low to strengthening convictions)
Another: actively trade a well understood core
Another: actively trade market / momo
Another: bungle around just jumping on tips
(I am “buy and shift”)
A concentration strategy: widely diverse port
Another: concentrate to highest of diverse convictions
Another: heavily concentrated into conviction (under 12)
(I am “heavily concentrated”)
Selection strategy: pick specific themes or industries or sec trends to ride
Another: deeply understand a concentrated core
Another: use a selection service or trading platform
Another: pick an open guru and follow blindly
(I am “deeply understood core” in “specific themes”)
I am not impressed with $FSLY's performance or mgmt ... yet again.
Rev +40%, decelerating -4pp YoY. That isn't the issue; they are usage based and bounce around.
But it's their first Q of SigSci -- adj out the $6M (an analyst had to wiggle out in Q&A), and they grew 30%.
Sorry, but for my rules, 30% rev growth is NOT hypergrowth.
Looking forward, guidance for Q1 was 32-36% YoY growth, but again adj out SigSci's expected 8M, and it's really 19-24% YoY. Terrible! Clearly no Compute@Edge contribution either.
Customer growth was +20% and enterprise custs slower at +12% (+3.5% seq). This is where I see the fracture in execution - you cannot scale if you aren't intaking new custs at a rapid rate.
Good news here: A new CRO was hired, hopefully to correct this.