I'm evaluating Nodejitsu for hosting Node.JS.
I'm trying to understand the relationship between Nodejitsu and IAAS providers. In particular is Nodejitsu providing me the machines or can I choose that my machine will be on some specific cloud provider in some specific zone.
Here's my need: My node service will need to communicate with a separate machine which I will set up myself (due to heavy customization). I want this machine to be on the same data center of my Node servers so in/out network cost and latency are minimized.
Per the nodejitsu FAQs, if you are on a personal plan, your host is Joyent's us-east-1 datacenter in North Virginia (USA). Business plan users are able to choose from additional datacenters which include Joyent's us-sw-1 in Las Vegas (USA), eu-ams-1 in Amsterdam (Netherlands) or Telefónica's eu-london-1 in London, England.
Looks like additional datacenters will include Joyent San Francisco (USA) and Telefónica in Madrid (Spain).
Update:
Nodejitsu recently experienced an extended outage which is documented in detail on their blog. A consequence of this was to drop support for Telefónica Instant Servers.
Related
Trying to look for messaging solution for the app that we are building and planning to host on Amazon. Is pubnub also hosted in Amazon?
PubNub Data Centers
Amazon EC2, IBM Softlayer, Rackspace and Microsoft Azure and plans for Google Cloud.
Where are clients connecting to when I deploy my application with PubNub?
PubNub is a globally distributed Publish/Subscribe communications Cloud. PubNub operates in multiple cloud vendor providers including Amazon EC2, IBM Softlayer, Rackspace and Microsoft Azure. Currently, PubNub is deployed in eighteen (18) EC2 availability zones across six (6) Amazon geographical regions:
EU Central
EU West
Asia Pacific
US West
US East
South America
Client connections are automatically routed and handled by the nearest data center for optimal response and throughput speeds.
Where are PubNub servers located?
What is the impact on my PubNub-powered application due to a data center failure?
I am talking about big companies, like ebay, Amazon. Do they host their application servers on Virtual Machine instances like EC2? Or they are still using physical servers? Is there a trend to move from physical servers to VMs because VMs are cost-effective and scalable?
Thanks!
Update: I found an article here: http://gigaom.com/2012/04/09/why-instagram-is-likely-moving-on-from-amazons-cloud/
Instagram was building on virtual infrastruture before, but its owner FB might be considering to move away from AWS.
But the question still remains: Are the applicaiton still running on virtualized OSes or on native OSes?
HostCabi.net surveys the top 500 hosting providers which accounts for the top 100,000 websites excluding the actual top 500 which are surveyed in the top 500 websites including who hosts them.
I'm not aware of a specific table that is going to show you what technology each hosting company provides, for that you would have to check on the website of the hosting company to see what service they offer.
AWS Direct Connect allows physical connections and BGP setup to connect AWS cloud with on-premise DC or customer's private cloud. I wonder if there's anything like this in MS Azure cloud. Not VPN based..
Kind-of.
It is announced as a future plan to support, but not yet accessible to customers. I believe this is what you are asking for: http://www.microsoft.com/en-us/news/press/2013/sep13/09-18msattpr.aspx
Hopefully we can see this announcements moving forward in the next year. Offering this solution to regions other than USA will also be interesting development of this solution, but currently nothing can be found on the internet.
As the cloud continues to be embraced by customers and partners around the world we are seeking to help them connect to their off-premise investments. Today, providers such as TW Telecom, AT&T, and Level 3 grant their clients the ability to connect to AWS or Windows Azure with scalable and flexible connectivity. Recently, Level 3 released ExpressRoute; a new service that allows customers to utilize a private connection from an Equinox Data Center to the Windows Azure Cloud. This service is similar in nature to the AWS direct connect service that was released a few years ago, providing a private network route to hosted cloud services that bypasses the public Internet.
ExpressRoute aims to reduce latency, and increase the speed in which clients can access their applications. To achieve this goal, Level 3 provides the client with 1 gigabit or 10 gigabit connection from an Equinox Data Center in San Jose to Windows Azure increasing your network throughput for large workloads. Today, Windows Azure does not offer multiple VPN’s into their cloud. To compensate for this issue, consolidating and connecting to the Azure cloud with a private connection ensures the transfer of your data.
Ref. from http://fastblue.com/cloud-connectivity-aws-direct-connect-and-windows-azure-express-route/
I gone through official announcement of General availability of Virtual Machine on Windows Azure. so i would like to move one of my client's production app on this. Requirement to set up this app on 2 different regions as below
Basically we have two demographics from Australia and Turkey from where people will be accessing the server, So we need two web servers one from Australia and another from Europe
- 2 Web servers with Window server 2008 r2
- Separate Database server with SQL server
- One Load balancer
Queries
1) What data centers are best suited for above regions?
2) we are not concerned about up-time instead we are concerned about performance at locale
based so to run application on 1 VM is okay?
3) In which region database should be hosted so both web server can access smoothly
regardless of performance issue. what should be region of it?
4) Can we setup load balancing such a way for one url it would transfer request to
Australian server and for other transfer request to European server?
5) I have calculated price with
2 VM (small) for to host window server
1 vm (medium) to host sql server
35 GB bandwidth
support for developer and it would cost me around $300,
Do I need to consider other pricing apart from this?
As for data-centers best suited, I suggest using
1. Australia - South-East Asia
2. Turkey - West Europe/North Europe
Now for the other question....
1. if you are looking at performance at locale...then you should look for setting up Windows Azure Traffic Manager (Performance-based)...this will give you best performance based on closest hosted service....
2. Windows Azure Traffic Manager (Round-robin-based) will serve the expectation of one request to Australia and second to Europe.
Note: you can set either of the above two but not both. It is different from load-balancing which works within a region (Read more...Azure VM load balancing vs Traffic Manager)
There are other cost like transaction cost, storage cost, etc....I suggest you read Microsoft Azure cost model for details. (Read more...http://blogs.msdn.com/b/windowsazurestorage/archive/2010/07/09/understanding-windows-azure-storage-billing-bandwidth-transactions-and-capacity.aspx)
Bear in mind, VMs can recycle...you can also explore cloud services (web roles) for your requirements.
I would suspect that you'd get more performance improvements by having one SQL database server and one web server that are in the same data centre, than have two web servers close to your customers both accessing the same DB. There is going to be more communication between the web servers and the DB than there will be between the clients and the web servers. You can then choose which set of customers you most want to keep happy and then choose the data centre close to them. If you want to keep your Australian customers happy, then you have another choice, because for cities in NSW that have access to the Southern Cross Cable, they get better response times from US data centres than they do from the asian ones.
We are on the process of choosing a Cloud provider to run our VMs, and we would like to better understand the differences between AWS and the new Azure.
Does anyone know if Azure Cloud (IaaS) have an equivalent for the AWS Route53 and Elastic Load Balancer?
If it has, how does it work? Do they also have several different zones across the globe capable of providing High-Availability with minimum impact on the latency?
Thank you
Microsoft doesn't have a DNS service like Route53. But does support common DNS operations such as c-name routing/forwarding to various Windows Azure services. As for zones, there are currently 8 locations that MSFT offers for hosting of Windows Azure, 2 in Europe, 2 in Asia, 4 in the US, and Fujistsu offers hosting of some Windows Azure services in Japan. With the possible exception of Fujitsu (I don't know the details of their SLA), all Windows Azure Services are at least 99.9% (the definition of "highly available" is subjective). Additionally, Microsoft has the Content Delivery Network with more than 24 nodes spread across the globe to help speed delivery and reduce the latency of content being delivered to almost any geography.
For the most part yes, there is equivalence. Windows Azure has always had a load balancer on any external endpoint (exposed over the VIP). It is not a super programmable firewall at this point (you cannot do more than simply remove yourself from rotation), but it works very well for web farm scenarios.
Cloud Traffic Manager works very well for geo-locating your services in Windows Azure. Just deploy your service in multiple datacenters, configure a few settings in traffic manager, and you are done. Most folks use it for 'best performance' (i.e. closest DC to customer), but also you can use it for 'failover'.
In terms of DNS, there is no equivalent service today for handling external DNS names. You handle DNS today with your own servers. That is one point where AWS is ahead.