Broadcom is again (see VA and Symantec purchases) chasing the train (hot property) long after it left the station.
General virtualization is already a commodity thanks to open sources like those you mention and others. VMWare should've innovated much faster/better in regards to containers, but I think they missed the boat (pun intended)
I recall that a few times I asked here about why when talking about virtualization, everyone defaulted to VMWare with Hyper-V coming second and only rarely mentioned QEMU or Xen, be it either professional environments or a home lab. It seems that people goes to the more popular Software available that works for them and rarely change. If this shakes things up, maybe I will see more people talking about the alternatives.
Funny that it goes directly against my own experiences here and every time home lab is mentioned 9 of 10 times KVM based ProxMox is recommended.
Veem Backup already supports Nutanix as well, and before you say Nutanix is expensive, consider that VMWare is already on the path to converting all clients to subscription-based by 2025. Source: CDW. After (or IF) Broadcom sell goes thru prices would go only one way.
You make it sound certain. Add price hikes like Atlassian and you are in gambling territory on the side of management, if this works out, or leads to serious cash-flow disruption. We've been a VMware shop since the GSX Server days. Saw EMC buy VMware, IPO, then Dell, now Broadcom. Every time this happens the company loses a bunch more greybeards who drove innovation and ran the show.
I get the appeal of these. Small, sometimes dirt cheap, low power, excellent to get your feet wet with clustering.
I agree that ECC is a must. If I'm not mistaken, DDR5 will not have the option to not use ECC, just the level of detection and recovery.
On VMWare subscription. As I've said I didn't read online somewhere on 4chan but heard it myself of a VMWare sales specialist working a major technology VAR - CDW. These are that person's words, not mine. As far as it would likely increase under Broadcom, that is a logical estimation.You make it sound certain. Add price hikes like Atlassian and you are in gambling territory on the side of management, if this works out, or leads to serious cash-flow disruption. We've been a VMware shop since the GSX Server days. Saw EMC buy VMware, IPO, then Dell, now Broadcom. Every time this happens the company loses a bunch more greybeards who drove innovation and ran the show.
I get the appeal of these. Small, sometimes dirt cheap, low power, excellent to get your feet wet with clustering.
But I would never recommend putting them into production. Can't expand them easily, castrated I/O capabilities due to TDP limitations, and worst, no ECC. Every administrator of IT systems amateur and pro alike should run rasdaemon and read
Why use ECC?
danluu.com
Unless you work for McKinsey or Accidenture where such errors create the wanted endless billable hours revenue stream. ;-)
Agreed.
No, the relationship between ECC and DDR5 has caused a lot of confusion.
Now that makes a lot more sense than what I was thinking it was about. I had a similar issue with ECC in HDDs where the on-disk ECC, DRAM cache ECC and interface ECC (or usually FEC?) were used interchangeably as "this drive has ECC". In a way you'd think that a development of DIMMS without a register but with ECC should be possible at some point. The whole non-ECC, ECC-but-without-a-register and ECC-with-a-register seems a bit segmentation for the sake of segmentation. Maybe there are actual silicon limitations where you can't produce NAND dies with ECC logic as cheaply as dies without it? Heck, if we can have ECC memory without a register, why bother making yet another series of SKU that don't have a register and also don't have ECC. Just ECC everything and let the host CPU decide
www.architecting.it
Funny because it was very predictable (and predicted) and then a bunch of new articles present it as super big news while all it really was is a "well duh" moment. Even their container offering is behind the curve as is their Kubernetes stuff. It seems they went down the complacency route and "keep existing customers happy" which seems to be the curse of companies that get a high acceptance/buy rate to then get stuck in. The only way to make more money at that point is keeping the people who are locked in already but making them pay more via a subscription.Maybe I've should've listened to this gent a bit more back in 2016:
Broadcom to acquire VMware as IT growth transitions to containerisation - Architecting IT
Broadcom is acquiring VMware in a transaction worth around $70 billion. This signals a transition away from server virtualisation growth.
p.s: @oneplane why funny?
fair enough. The reason I posted it as re-enforce my own other sources on the fact VMWare was already on the path to the subscription model, but someone above didn't quite believe me on that.
I bet that the free ESXi is definitely on the chopping block. Their intentions aren't to expand the market base, to milk out as much as possible from existing customers. Free ESXi goes very much against that idea.
