Esta es la primera de una serie de publicaciones en el blog sobre el esfuerzo que estamos llamando Linode: NextGen. En los próximos días le informaremos de otras mejoras y cambios, pero hoy queremos informarle de las actualizaciones de la red.
Vamos a gastar un millón de dólares en hacer nuestra red más rápida. Mucho más rápida. Routers Cisco Nexus 7000. Conmutadores Cisco Nexus 5000 con extensores de tejido Nexus 2000. Linode El límite de la red saliente se ha multiplicado por 5. La cuota de transferencia mensual saliente se ha multiplicado por 10.
Estamos actualizando toda nuestra red, en los seis centros de datos. Se está sustituyendo todo, desde los routers de alto nivel hasta los conmutadores a los que se conectan las máquinas anfitrionas. Hemos podido hacerlo sin ningún tiempo de inactividad, y esperamos completar las actualizaciones en las próximas semanas.
¿Por qué hacer todo esto? Sencillamente, nuestra antigua arquitectura de red se ha quedado pequeña y necesitábamos algo que pudiera:
- lidiar con nuestra escala;
- mejorar en gran medida el rendimiento;
- reducir la latencia; y
- añadir redundancia en la capa de acceso/host
Elegimos los routers de la serie Nexus 7000 de Cisco porque son los únicos dispositivos que ofrece Cisco que tienen tablas de direcciones MAC y de enrutamiento lo suficientemente grandes como para satisfacer nuestras necesidades. Para nuestros nuevos switches de capa de acceso, nos decidimos por los switches Cisco Nexus de la serie 5000 con los Fabric Extenders de la serie Nexus 2000 para la conectividad de la parte superior del rack.
La totalidad de nuestra pila de redes en cada centro de datos tiene redundancia incorporada, de arriba a abajo. Cada host está conectado a dos extensores de tejido separados con enlaces gigabit enlazados. Esto no sólo proporciona redundancia en caso de fallo de una NIC, sino que el tráfico del host se equilibra en sus NIC.
Cada extensor de tejido está conectado a un par de conmutadores Cisco Nexus 5000 mediante enlaces de 10 Gbps, equilibrando la carga de su tráfico en cada conmutador. El equilibrio de carga significa que el rendimiento efectivo entre el par de extensores de tejido y el par de conmutadores es de 40 Gbps.
Cada par de conmutadores 5k está conectado de forma similar a nuestros dos routers en el centro de datos: cuatro enlaces de 10 Gbps entre ellos, con equilibrio de carga, de nuevo con un total de 40 Gbps de rendimiento. Por último, cada router dispone de varios enlaces de 10 Gbps con nuestro proveedor de acceso, también con capacidad de conmutación por error y equilibrio de carga.
Utilizamos los mejores equipos de conmutación que vende Cisco. La mayor capacidad de la red aumenta en gran medida nuestra resistencia a los ataques DoS, y la configuración de carga equilibrada y redundante significa que los fallos de hardware no afectarán a la conectividad.
También hemos ampliado el límite de red saliente por defecto de Linode de 50 Mbits a 250 Mbits/seg. Todo lo que necesitas es reiniciar para recibir este aumento de cinco veces. Recuerda que este es un límite para el tráfico saliente solamente. El entrante no está limitado.
Por si fuera poco, hemos aumentado en un 1.000% la cantidad de transferencia saliente incluida en todos los planes. Así es, ¡10 veces la transferencia incluida!
- Linode 512 actualizado de 200GB a 2000GB (2TB)
- Linode 1G se ha actualizado de 400GB a 4000GB (4TB)
- Linode 2G se ha actualizado de 800GB a 8000GB (8TB)
- Linode 4G se actualiza de 1600GB a 16000GB (16TB)
- Linode 8G se ha actualizado de 2000GB a 20000GB (20TB)
Como antes, todas las transferencias entrantes son gratuitas.
En los próximos días os hablaremos de las partes 2 y 3 de Linode: NextGen. Creemos que las encontrarás igual de emocionantes. Que lo disfrutes.
Comentarios (141)
Wow, this made my day. Low transfer caps was my only gripe with Linode before, but not now!
Yes!!
I have been hoping that Linode would do something like this soon. Linode has sort of been stagnating lately, but I had a feeling you guys were working on something big. THANKS!
oh my god
•Linode 512 upgraded from 200GB to 2000GB (2TB) WOW
thank you 😀
Linode > *
Motherofgod.ico
10x increase in allocation AND 5x increase in speed!? Linode, you just blew my mind.
Awesome! WowwwwwwwwwwwWWWWWWW!!!!!!!!!!
thank you 🙂
Wild!
Wow. Thank you!
10 times? Really! Amazing!!
Great news 🙂 thanks
Its times like this I really know choosing linode as my host was the best idea ever 🙂 Thanks a lot for this upgrade!!
Makes my transfer usage bar when I login to my linode look even worse:
http://i.imgur.com/1an4auM.png
Damn. My ISP could learn from you.
I can’t tell you how happy I am about this, I was just about to cancel my account in favour of a company offering more transfer. I was filling in the cancellation form when I saw the Facebook post, so glad I spotted this in time!
This is exciting news. I’m glad to see Linode isn’t resting on their laurels.
I’m crossing my fingers that one of the future announcements you mention has to do with storage, and particularly I/O performance. I want to move the rest of my company’s servers to Linode, but for a couple of applications in particular, we need faster I/O than the nodes we tested could provide.
That said, your product is fantastic, and I never hesitate to recommend Linode to other people. Keep up the great work.
Wow,thank you linode!
I’m surprised that nobody mentioned that based on the diagram you run all of Linode on 6 hosts. Amazing job. 😉
The best this is: I trust linode and I know they can deliver what’s promised.
Other providers may offer stuff like this, but I doubt they won’t kick you out as soon as you start hitting those limits.
I’ve been a costumer since Jan/2009 and I look forward being a costumer for years to come.
I wonder why I was wasting my time and money to other hosting providers…
One more thanks from me Linode!
Awesome! We love Linode
My February cap was almost exceeded (96% of my 200GB quota used), so this could hardly have come at a more opportune time. Thanks, Linode !
Wow! This is why I’ve been sticking with you guys, there is nobody else that offers this type of service and constantly increasing features. Proud to be your customer 🙂
Wow, signed up with you guys moving over from dreamhost and it has been a dream – so happy about all of this. thank you!
Oh well there goes up my uptime! At least it’s for an awesome reason. Thank you Linode
Thank you guys so much! Very exciting news, can’t wait for the other NextGen stuff.
Amazing Linode! Thank you very much!
Thanks, It’s wonderful.
Hi Linode,
please add the ability to provision bare-metal servers with fast disks attached. With such an hybrid setup you are going to dominate the market if you move FAST.
WOW, this is amazing Linode! Thanks for this!
Sick! Thank you Linode.
This is super news! Thank you very much!
“All you need to do is to reboot to receive this five-fold increase.”
So… when can we expect these changes take effect? We need to schedule our downtime for those reboots. Yummy, yummy bandwidth… 🙂
Well congratulations linode, you just won my business.
You may want to double-check with your Cisco support engineer about those FEX-to-5k cross connects. They generally recommend the FEX’s connect to a single 5K. Also, the VCD configuration on the a pair of 5Ks can be a little wonky.
Best of luck, looking good!
Wow, this is really nice! Amazing 🙂 Good job Linode.
Peter:
We are actually using Enhanced vPC with our 5Ks and FEXes, which is a Cisco-supported topology. It allows us to connect each FEX to two 5Ks, as well as have both 5k-to-FEX vPCs and FEX-to-host vPCs.
We’ve been using this topology since mid-2012 and it’s been working great for us. 🙂
Thanks guys! Always had great performance and uptime with linode, and this is a very nice gensture.
I am so happy I chose linode.
Das is uber guten.
Damn glad to be a Linode customer. Thanks and keep up the great work!
Wow. Thank you.
Good to be a Linoder
Oh! Will be opening my account again then!
Awesome news!
Awesome, thanks for the additional xfer! As mentioned, hope it’s in preparation for something like an OpenStack Storage implementation for a CDN, err cloud offering but will also allow for creation of large (inexpensive) persistent and ephemeral disks for their nodes similar to EC2 and Google Compute volumes.
This is incredible! You’ve just put money back in my pocket. It is totally unheard of for a host to do this. You have a customer for life.
Once again I become very confused when people recommend anything but Linode for a VPS
Linode, model VPS ever!
Marvelous. A great move by Linode.
Linode Nextgen Rocks 🙂
Linode is a no-brainer now.
I won’t get sold that easy… The truth us, bandwidth consumption is possible only when you offer extra cpu and extra memory… Without that, I bet you could offer 100x bandwidth and still no difference in the overall delivery.
reaaaally great news Linode! Thanks a lot
The homepage shows the blog post as “March 2012” 😛
@ cralor – fixed, thanks.
> Linode 512 upgraded from 200GB to 2000GB (2TB)
Fantastic, great to see you passing improvements down to customers. I knew you guys had a good reputation (which is why I signed up with you) and now I’ve experienced why first-hand, brilliant.
<3
Linode you are awesome!!!!!!
I recently scaled back my linode to a 512 and moved the main site to a Dutch VPS hoster that gave me 1GB ram for 1/2 of the 512 price. I loved linode, but its getting too expensive for hosting rails apps that eat memory.
here’s one hoping for mem increasings, the service of Linode is unparalleld.
*jaw drops*
Way to go Linode!
Interesting. Of course my linode64^Wlinode512 rarely gets to 5% of my current bandwidth 🙂
I wonder if this new design will allow for FCoE or iSCSI based SAN environments in each datacenter. That could increase resilliency (atlanta47 dead? Migrate to atlanta53 with a flick of a switch!) and might let you scale disk storage (a repetetive ask in the forums; again not something my small needs care too much about).
Interesting, indeed!
Great news, Linode latency was already away better compared to my DigitalOcean VMs. Now the 4x price tag is starting to look justified.
Really stunning. We will move everything from ec2 to linode next week.
Wo Wo Wo
wow, 10x ! Perfect!!
Is overage still $0.10 a GB or has that been lowered as well?
WOW! The network is quite fast now!
Great news! I’m glad to see that there are some big changes happening at Linode. Can’t wait for the other articles.
Incredible changes, 10x! It’s awesome!
Cool
oh, my aws will be discarded.
Pretty cool!
Was wondering just yesterday, when do we hear about whats inside the big boxes that were delivered last month …
now need to wonder, what do i do with this extra bandwidth..
“wonder”ful linode..
Wow. Thank you!, very good!
Wow, I just made the move to Linode two days ago. I wonder what’s in store with Parts 2 and 3.
-jaw drop-
Thnx Linode
*pats* I loves you linode *squeeze*
“All you need to do is to reboot to receive this five-fold increase.”
“In the next few days we’ll tell you about parts 2 and 3 of Linode: NextGen.”
So, should I wait for parts 2 and 3 for the rest of the goodies to come out before I reboot my node? 🙂
Good news , but I can not use more than 10% of my transfer 🙂
so,amazing!!! love all your guys so much!!
Great work guys!
You guys are doing great. Great service backed up by awesome updates!
I’m curious to see what happens after I reboot 🙂
WOW!!!
SO F***ING awesome!!!!!
Great Job!
Great ! 200 to 2000
Wow!!! So proud of Cisco Nexus gear. Have you talked to any service providers using it???
It’s buggy at best. Expect a random change to cause a spanning tree loop, which will take the entire network off line.
The saying used to be “no one gets fired for buying Cisco”. That was before Nexus. We deployed it and it killed our entire networkore than once. We were wishing we would have gone with something else…Juniper???
Shit! That’s fantastic!!!
Staggering upgrade for customers – thanks Chris
I was contemplating moving to a competitor and using an S3 bucket… not now, great work 🙂
That is just great news. Keep going!!!
Just guessing the Part 2 and Part 3… It is going to be RAM, Last RAM upgrades where in 2007 and 2010.. its 2013, so expect RAM to upgraded.. and as Storage was upgraded recently, their wont be much in it.. Additional functionalities and better Load Balancer..
I just wish I could run FreeBSD on Linode.
machbio, One of them will have something to do with the ability to have accounts without a linode as mentioned (rather hinted at) earlier. Only thing I can figure is that is a new service beyond the linodes. Otherwise no one would really have a need.
The other is probably an upgrade to all resources (ram,cpu, and disk).
Just guessing though
Linode, this and stability is why your customers are so loyal.
Fingers crossed for memory upgrades….
You rock!
Also, unrelated, would you consider spanning the private networks across datacentres?
DR for Linodes would be awesome.
I would love to see memory and increased io/shared disk
Also, love Linode, they always impress me!
Awesome!!
Great Job!!
Cisco Nexus 7000 routers? I only know Cisco Nexus 7000 series switches. And why chose Nexus? I think products of Juniper is better for High-End networks.
Thanks. extra memory will be very nice!
I still don’t see FreeBSD support, so what’s the point?
+1 wow
Well, THAT’s a nice day-early birthday present. Thanks!
Pedantic mode: isn’t that a 900% increase, for a new total of 1000% of the original? Either way, pretty awesome!
Bob, can’t you do FreeBSD using pvgrub?
Wow, thanks!
WP engine customer saying thanks and thinking it’s been a benefit me to. I have a friend that wants to upgrade shared hosting to something faster I’ve given him some decent alternatives and I’m wondering if anyone has any great ideas for under 20 bucks that somebody used to shared hosting could run?
Wow.
I love you, Linode.
Wow! Very nice of you to do this for us. This reminds me of why I love Linode so much.
@Shane (3/7)
Yea… Mine looks nearly as bad.
http://puu.sh/2eXtV
I logged in to my manager and my jaw hit the floor when I seen the transfer. Awesome!!
Linode! Can you please tell me why you are so awesome? Eagerly waiting for the next piece of the cake 😀
Thanks a lot!
Thanks for the upgrade!!
Hi,
Excellent upgrade! I’m looking forward to the next surprises!
Blake
This is really great! Thanks for the extraordinary Support Team who are always on top of the Support Tickets.
Linode always has something to offer every other quarter. Part 1 of NextGen is awesome!
What will be even more awesome in NextGen Part 2 & 3?
– Bigger Plans with More Cores/CPUs (today its at 4)?
– SSD Drives for better IO?
– More RAM?
– Built in Firewall?
– Elastic IP / Global Load Balancer across the DCs?
– and may be pre-configured MySQL like DBs for HA n Replication
Would be really great to see Digital Ocean like hardware specs/configurations – specifically CPUs & SSD drives.
Cool! I wish they also increased the amount of RAM included in each plan 🙂
Awesome! Love Linode.
This is great – well done linode!
I’m running the exact same configuration in my data centers and it’s been very reliable. Just make sure you pick a good code base and have all the EPLD updates done before you go into production.
I’d be interested in learning what you’re using for a border router.
Thanks!
woah! Great! Linode is really always on top of the line. keep it up!
Gonna be very hard filling up this bar 🙂 http://i.imgur.com/1bazgLG.png
Solid improvements Linode, very happy with service so far… keep up the good work and you’ve got a customer for life!
Thanks,
Paul 🙂
I guess the next part maybe upgrading the hard drive to hybrid type (HDD+SSD).
Nice move! Great thanks folks! 🙂
When part 2 & 3 would come??
You guys are the BEST! Thanks!!
i bought 512mb ram plan for one year. using it almost for one month. everything was good until yesterday. starting to saw sudden crashes and server totally stops responding until initiate reboot.
i was just thinking to move mediatemple vps because they are offering 2x more ram with same price but unfortunately their servers are located only in us.
also saw the commercials for digitalocean ssd for storage and again looks cheaper but their promises looks very unrealistic to me. i think i should wait and see how serious they are.
in a nutshell, i would love to see some RAM upgrade and probably bandwith was not the main concern for majority of users.
Linode rocks.
I’ll have to second the opinions of many commenters here and say that RAM is probably more of a concern for many of us (including myself!).
I’m having to spend ‘extras’ each month just so that I can configure a good server, I hope to see RAM increased in the near future to match the competitors.
In my eyes this is the only think that’s niggling me at the moment with Linode…
Thanks,
Paul 🙂
You are the best! Thank you!
Out of curiosity, do Linode staff actually read these comments? Paul 🙂
Paul: Yup.
Crossing my fingers, hoping the next free upgrade involves storage 😀
Go Linode!
I never understood the RAM and Storage ‘extras’ from a customer view. It is cheaper to upgrade to a larger Linode. It requires a reboot/resize either way.
Great work linode! You guys are great!
Great news ! thanks a lot 🙂
Linode, this is insanely sick!!!! You just equipped superman with the Iron Man Suit. UNBEATABLE!!!
I rarely have to log in to linode as the service is soo rock solid I rarely need to.
However I was actually checking as I am going to change how I use my VPS and feared my Linode will not be able to handle it… I am blown away by your free upgrade (which will indeed handle what I want to do) and all I can say is WOW! Thank you, I get to stay!
Amazing!!, that´s why I moved all from GoDaddy to Linode…you rock guys!!!
Linode is the only serious player in the market, you certainly beat everybody else. Thanks!
Linode Team – Just sending you guys some love – you’re so awesome.
Thanks for such a great service – you’re an example to others of how to get it right….
Wow, Made my day. Love it Thank you guys for being awesome 🙂
@Dash, I looked at the offerings on DO, Ramnode, SSDVPS, lowendbox..
These folks are running RAID10 on Intel 520 SSDs (consumer grade SSD or equivalent), I don’t think that’s a very good idea, I don’t even run RAID0 on my HDD setup (I run RAID1).
The prices there are rock bottom but I guess I’ll find out why when an entire host goes down from multiple SSD failures (If I hosted with them I’d be sharing a VPS with write-happy people), and I have to tell my client has been down for 3 days because I can’t get ahold of anybody.
They also seem like they’re almost fly-by-night, but at least DO seems to be totally domestic and has photos of their peeps.
SSD for caching I like, such as with ZFS l2arc; but as the main drive, and shared with other people? NOOOOOOOO, just mho