IPv6 still 5–10 years away from mainstream use, but K8s networking and multi-cloud are now real
- Reference: 1626332692
- News link: https://www.theregister.co.uk/2021/07/15/ipv6_istilli_510_years_away/
- Source link:
So says the firm’s 2021 Hype Cycle for Enterprise Networking, published last week and now grasped in The Register ’s claws.
Let’s start with the graphic depicting the Hype Cycle in all its glory.
[1]
Gartner Hype Cycle
for Enterprise Networking 2021. Click to enlarge
While IPv6 kept its status – thanks to IPv4 remaining adequate and the benefits of migration rated “low” – Gartner added six new items to the cycle, namely:
6G – While still theoretical and unlikely to be deployable before 2030, Gartner advises keeping an eye on developments;
AIOps – What’s not to like about AI applied to network monitoring and management, especially if it delivers on the promise to dig up issues and then automatically fix them before humans even notice the problem;
eBPF – The Extended Berkeley Packet Filter (eBPF) is an enhancement to the Linux operating system kernel that allows specific instruction sets to run (sandboxed) inside the kernel. Likely to become a big part of Linux distros because it improves the reliability and security of applications;
NaaS – Vendors want their slice of the network-as-a-service market, but Gartner warns buying one surrenders control, may not be cheaper than other procurement models, and delivers few benefits;
Network observability – Included to help buyers understand that this is essentially a rebranding of “network monitoring” and vendors are trying to pass it off as something new, which it isn’t;
Private 5G – replaces private standard LTE and rated as one to watch as it will deliver campus wireless more cheaply than WiFi. Also an enabler of edge computing, factory digital twins, and edge AI for computer vision.
Other elements of the document The Register found notable include intent-based networking dismissed as something you don’t need to worry about for a decade. Cisco may have gone all-in on the concept, but Gartner rated its efforts and those of other vendors as cursory.
“Nearly all products that are marketed by vendors as ‘intent-based’ fall short of the full capabilities of an intent-based networking system,” the Hype Cycle states.
[2]
Software-defined networking was again rated obsolete, because there are no real products in the field. “While technologically obsolete, SDN terminology is widely used by vendor marketing efforts,” the document states. “This remains the top driver behind SDN discussions today.”
[3]
[4]
Zero-trust network access (ZTNA) is real, the document states, because it “yields immediate benefits by shielding services from attackers”.
“In contrast to basic VPN products, ZTNA removes full network access and improves user experience, flexibility and adaptability,” it adds. Gartner rates the technology as “adolescent” but sees it reaching the plateau of productivity in two to five years.
[5]
Another big mover is internet-only WAN access, which Gartner reckons will be used by 30 per cent of you by 2023 – double the rate of 2020.
Networking stacks that span multiple clouds are also a boom market, with the analyst predicting 35 per cent adoption among multi-cloud users by 2025, compared to around three per cent today.
Secure web gateways and cloud access service brokers are also on the march, as are branch office firewall as a service.
[6]Desktop PC sales bounce back – but only because of laptop component shortages
[7]Dedicated (Local) Cloud Infrastructure-as-a-Service to grow almost 1000 per cent in five years
[8]The world has a plastics shortage, and PC makers may be responding with a little greenwashing
[9]Pyjama bottoms crew, listen up: In 2022 we'll still be at home
400Gbps Ethernet is rated a niche concern, other than for hyperscalers and those who run AI/ML applications in-house. Gartner rates it as the likely successor to 100Gbps Ethernet, despite work having commenced on 800Gbps Ethernet.
Gartner considers Huawei a vendor worth considering for 400Gbps Ethernet, as well as for WiFi 6 and Private 5G networks.
[10]
The analyst firm thinks five networking techs have hit the peak of the hype cycle. Software-defined cloud interconnect is one, partly because its main providers are obscure companies. Kubernetes cloud networking is another, because it’s been recognised that Kubernetes’s native networking features aren’t up to the job.
5G is established, and Secure access service edge (SASE) – a term Gartner coined – has been recognised as a quicker and simpler way to secure the enterprise.
Multi-cloud networking is also considered a tech whose time has come. ®
Get our [11]Tech Resources
[1] https://regmedia.co.uk/2021/07/15/screenshot_gartner_hype_cycle_for_enterprise_networking_2021.jpg
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/networks&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YPAHPHg2ZOr2PuE2Eq6fjQAAAAw&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/networks&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YPAHPHg2ZOr2PuE2Eq6fjQAAAAw&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/networks&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YPAHPHg2ZOr2PuE2Eq6fjQAAAAw&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/networks&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YPAHPHg2ZOr2PuE2Eq6fjQAAAAw&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2021/07/12/idc_worldwide_quarterly_pc_tracker_q2_2021/
[7] https://www.theregister.com/2021/07/06/idc_dciaas_prediction/
[8] https://www.theregister.com/2021/06/28/recycled_plastics_in_tech/
[9] https://www.theregister.com/2021/06/23/gartner_wfh_report/
[10] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/networks&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YPAHPHg2ZOr2PuE2Eq6fjQAAAAw&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[11] https://whitepapers.theregister.com/
Re: Is this the most sensible Gardner report ever?
IPv6 has lost the war. It will never see serious use. For the average sysadmin punter, what are the real, day to day problems that IPv4 has? The biggest by far is publicly routable IPv4 exhaustion. Solve that one, with backwards compatibility, and you have a winner.
Of course, incentivising/forcing organisations with Class A and B addresses that they use internally and never see the light of day, to release those addresses, is probably going to be a better short term winner.
Re: Is this the most sensible Gardner report ever?
“ IPv6 has lost the war. It will never see serious use.”
I largely concur but ipv6 is being used seriously by some, iirc some mobile carriers are ipv6 native.
It’s difficult when only some sites and users are ipv6 and some mechanism is needed to choose between the 2.
I’m on VM and Apples iCloud private relay will show me as having an ipv4 & ipv6 address (private relays job is to obfuscate my real ip), clicking google search links often fail unless I search from ipv4.google.com.
Security wise it’s probably a good idea to use ipv6 internally with ipv4 only internet connections. It’s an additional layer miscreants need to hop through!
Re: Is this the most sensible Gardner report ever?
The only use for IPV6 is large infrastructure, i.e. Telco. "The internet" is what Average Joe uses. That is mostly an IPV4 affair, and will remain so as long as that's where the majority of Average Joes hang out. Anyone creating an IPV6 only website will be shooting themselves in the foot commercially because most of their customers are IPV4 only.
Re: Is this the most sensible Gardner report ever?
Anyone creating an IPV6 only website will be shooting themselves in the foot commercially because most of their customers are IPV4 only
Given that you have have thousands of websites on a server cluster behind a single IPv4 address there is no particular reason to ever end up in this situation.
Re: Is this the most sensible Gardner report ever?
I have a cheap VPS which only supports IPv6 but you can still host a website that is accessible from IP4 only clients using Cloudflare proxying. You don't even need to have a paid Cloudflare plan its included in their free tier.
Re: Is this the most sensible Gardner report ever?
Decades on, I still think that the winning solution would have been just to add another two numbers to the scheme used by IPv4 and leave everything else alone. It'd easily have hit 100% utilisation by now because everybody would have switched to it without hesitation.
This would take you from 4.1 billion addresses to two hundred sixty eight point eight trillion addresses, with the point eight at the end being two hundred times larger than the existing IPv4 address block.
Re: Is this the most sensible Gardner report ever?
And, decades on, I shouldn't have to be repeating this...
It really doesn't matter how many digits you add. Simply adding a single bit makes half the potential new address space unreachable to "legacy" systems and that moves you no further forward unless you have a credible plan EITHER for ongoing interoperability (using some form of address translation) OR to migrate the legacy systems before the additional address space is required.
It was envisaged with IPv6 that the revised protocol stacks would be in place everywhere before the address space ran out (courtesy of CIDR-based life-extension) and indeed that largely happened in terms of end systems (most of which have been IPv6-capable for years) - and in core networks - but retail ISPs largely stuck with what they knew. I banged on about interoperability for some time and gave up because there was a complete lack of mutual understanding between people designing protocols and people deploying networks (faults on both sides). So without either interoperability or prompt migration, IPv6 uptake has been dismal.
But that would have been true however many extra digits the addresses had.
Re: Is this the most sensible Gardner report ever?
Sure, but all you had to do to migrate to the new standard was allocate a bit more memory for the address, I think people would have done it by now. IPv6 requires you to completely redesign your network topology.
Re: Is this the most sensible Gardner report ever?
"It will never see serious use" - 33% of traffic to Google is IPv6. https://www.google.com/intl/en/ipv6/statistics.html
Re: Is this the most sensible Gardner report ever?
I bet the vast majority of that is mobile phones.
Re: Is this the most sensible Gardner report ever?
No, because I don't see a 25 year old technology suddenly becoming popular in the next 5-10 years.
25 years in computing terms is ancient, obsolete even.
The "problem" that IPv6 aims to solve is the use of NAT at network gateways. The thing is that most people want to have NAT at their gateway and consider it to be a good thing.
IPv6 fanbois will tell you that NAT is not a security feature and shouldn't be used as such. They are wrong. Obviously it shouldn't be your only security measure, but it does help, a lot.
These aforementioned fanbois will also tell you that there is another, more complicated way to achieve the same thing in IPv6. Sure, but it is more complicated, not as well tested, and offers no practical benefits over IPv4.
IPv6 does offer more numbers to go round, and more numbers is good. Why can't we have an IPv4.1 or something that does the more numbers thing, but leaves everything else the same?
Version 6
IPv6 is already quite popular in use, but not at many cloud vendors. Cloud vendors have been hoarding IPv4 addresses and it's in their best interest that you feel like you need one. It's also wrong to look at your workplace's network or Docker and assume that their IPv6 isn't working. Except for some old telcos that never upgraded past DSL, most US Internet connections are IPv6.
Cellular 6G is completely dead until it can find a reason to exist. It's current ideas expand upon all the features of 5G that are proving to be unpopular and unpractical. A mmWave access point every 100 meters, an edge compute datacenter every 1km, and a world of unified AI with perfect connectivity? Maybe someday, but not with this tech.
Is this the most sensible Gardner report ever?
For the first time ever I find myself agreeing with most of that.
Maybe I’m ready for management after all?
SDN is just companies flogging stuff that sits in the way and does vxlan.
Private 5G & 5G will be a game changer
ipv6 has little benefit for most people
Intent based networking is a noble effort but Zero trust is the way forward.
I personally feel private wan circuits are more secure than wan over internet but zero trust is there for that.