Hi All,

I am trying to source an Enterprise Cloud / IaaS provider in Canada. By "enterprise" I mean offering services like:

- HA instances (e.g. Vmotion in VMWare)
- SSL Offloading / Load Balancing
- Managed Firewall
- Monitoring / Alerting / System Graphing
- Snapshots / Backup with Offsite Vault in retention policy

There are a bunch of "Slicehost-like" providers in Canada like Stardothosting but that's not what I'm looking for.

There have been a couple threads on this and someone did mention TELUS. I have talked to them, they are expensive and do not offer Monitoring / Alerting / System Graphing in any meaningful way. Others I have talked to like Radiant and Primustel have also come up short. I am pretty deep in discussions with Tenzing but they just threw me a curveball late in the game and I'd like to explore any other providers that you might recommend.

TIA

Cheers

I believe iWeb is the only provider that offers something close to your require.

Is there a specific reason why you want to limit this to Canada? In an IaaS environment there's rarely a benefit to such a limitation. We have quite a few IaaS customers from Canada for our parent company and there's cost savings, increased bandwidth and Data Center capabilities, as well as other benefits which typically outweigh that of Canadian providers according to our customers.

So just some food for thought...

Is there a specific reason why you want to limit this to Canada?

The threat of the Patriot Act (real or perceived) disqualifies the US.

Thanks for your reply.

Cheers

I believe iWeb is the only provider that offers something close to your require.

Had a quick chat with them, IMHO their offering is Slicehost-like.

Thanks for your reply.

Cheers

The threat of the Patriot Act (real or perceived) disqualifies the US.

Thanks for your reply.

Cheers

What exactly about it?

What exactly about it?

It's unacceptable to our customers that their data reside on US soil. It's an requirement like any other, the Canadian government calls the threat "remote": http://www.tbs-sct.gc.ca/pubs_pol/gospubs/tbm_128/usapa/faq-eng.asp#Q3 but I have to respect the concerns of our customers.

Cheers

Gotcha. If any of the data happens to pass through the US anyways (assuming any of these customers have any users that are in the US) it's still subject to it, but interesting to hear such feedback. Thanks! :)

I dug up Bell: http://www.bell.ca/enterprise/EntPrd_Virtualization.page and Scalar: http://blog.scalar.ca/cloud-practice/ .

I expect the Bell offering is comparable to TELUS in being overpriced and underserviced.

Anyone heard of Scalar? They look like a good candidate.

Cheers

I am facing the exact same situation you were. I was wondering where you ended up? What vendor did you choose? How have they performed for you?

We are about to go live with Cartika's cloud solution. I don't know if they have all the features you are looking for (some of those were not requirements for us), but they are in Canada and seem to be reliable so far. Their cloud is in Peer1 in Scarborough. We have the Canadian location requirement for some customers also.

Is there a specific reason why you want to limit this to Canada? In an IaaS environment there's rarely a benefit to such a limitation. We have quite a few IaaS customers from Canada for our parent company and there's cost savings, increased bandwidth and Data Center capabilities, as well as other benefits which typically outweigh that of Canadian providers according to our customers.

So just some food for thought...

I have been barking up this tree for years - and at the end of the day, we opened a Canadian facility to address this. Trying to explain to Canadian companies that their risk was the exact same with data in the US or CAD was pointless. Canadian customers want data residing in Canada - this is the bottom line. The patriot act scares every serious enterprise customer in Canada away from the US. As a Canadian based company who previously had data only in the US, we adapted our strategy to meet the needs of our customers.

Having said this - datacenter capabilities and pricing and other benefits are no different between the US and CAD - we do not charge any differently for services in CAD or the US. It is not appropriate to even suggest such things. Canada will one day be one of the world leaders in Datacenters - our natural cool climate, political stability and extremely strong economy as compared to the US, make us an ideal candidate for eco-friendly, efficient, advanced and politically stable data centers required moving forward.

Hi All,

I am trying to source an Enterprise Cloud / IaaS provider in Canada. By "enterprise" I mean offering services like:

- HA instances (e.g. Vmotion in VMWare)
- SSL Offloading / Load Balancing
- Managed Firewall
- Monitoring / Alerting / System Graphing
- Snapshots / Backup with Offsite Vault in retention policy

There are a bunch of "Slicehost-like" providers in Canada like Stardothosting but that's not what I'm looking for.

There have been a couple threads on this and someone did mention TELUS. I have talked to them, they are expensive and do not offer Monitoring / Alerting / System Graphing in any meaningful way. Others I have talked to like Radiant and Primustel have also come up short. I am pretty deep in discussions with Tenzing but they just threw me a curveball late in the game and I'd like to explore any other providers that you might recommend.

TIA

Cheers

Did you look at https://www.zunicore.com ? I am pretty sure their infrastructure is in Canada.

:)
D

Canada will one day be one of the world leaders in Datacenters - our natural cool climate, political stability and extremely strong economy as compared to the US, make us an ideal candidate for eco-friendly, efficient, advanced and politically stable data centers required moving forward.

But wouldn't datacenter growth be affected heavily by the network effect? For example, if all the providers seem to have a POP in Ashburn, VA, then it makes sense more datacenters will start up in that area because everyone else they want to connect to is already there.

If that wasn't the case, then why aren't there more datacenters in areas with cheap power but outside major cities? I know some really large providers who operate their datacenter internal use (Google, Facebook, etc) do this, but it doesn't appear like Equinix (for example) is doing so.

On a per-capita basis, how does square footage of datacenter space per person compare between the US and Canada?

Basically, I was under the impression that the reason the US has a lot of datacenters is that a lot of the internet and companies surrounding that technology originated there (Virginia and San Jose), and then the dot-com boom resulted in huge amounts of infrastructure being built out. Then add in the network effects, and that seals the deal.

Staying slightly on-topic, I agree that a number of Canadian customers simply insist their data reside in Canada, even if for practical reasons it's just as safe in the US.

But wouldn't datacenter growth be affected heavily by the network effect? For example, if all the providers seem to have a POP in Ashburn, VA, then it makes sense more datacenters will start up in that area because everyone else they want to connect to is already there.

If that wasn't the case, then why aren't there more datacenters in areas with cheap power but outside major cities? I know some really large providers who operate their datacenter internal use (Google, Facebook, etc) do this, but it doesn't appear like Equinix (for example) is doing so.

On a per-capita basis, how does square footage of datacenter space per person compare between the US and Canada?

Basically, I was under the impression that the reason the US has a lot of datacenters is that a lot of the internet and companies surrounding that technology originated there (Virginia and San Jose), and then the dot-com boom resulted in huge amounts of infrastructure being built out. Then add in the network effects, and that seals the deal.


Absolutely everything you said is correct. Having said this, nothing stops density from increasing somewhere new. Now that PEER1 has built out to Scarborough and has multiple providers with fiber onsite, I imagine this particular industrial park has the potential to be a major data center hub for the Toronto area.

its why I said Canada will one day be... vs Canada already is :)

Canada will one day be one of the world leaders in Datacenters -

The driving force moving forward for building data centers will not necessarily be the same as they historically have been.

Iceland right now is the trendy place to build out eco friendly data centers. I imagine Northern Canada will eventually be a target as well, especially considering the investment the federal government has already made in massive fiber pipes through the northern part of the country.

You can find some cool information on iceland datacenters HERE (http://www.google.com/search?gcx=c&sourceid=chrome&ie=UTF-8&q=google+iceland#sclient=psy-ab&hl=en&biw=1694&bih=876&source=hp&q=data+center+iceland&pbx=1&oq=data+center+iceland&aq=f&aqi=g1g-m2g-b1&aql=1&gs_sm=e&gs_upl=11875l14684l0l14828l19l12l0l1l1l0l439l2849l0.3.5.2.1l12l0&bav=on.2,or.r_gc.r_pw.,cf.osb&fp=f968ab45d699d19)

facebook has built a data center in Sweden, google in finland I think - heres another cool link with some of these stories

http://www.greendatacenternews.org/topic/connection/iceland/google/

In the future (actually the future is now I think) - people are beginning to value more efficient/eco friendly data centers (which typically require cooler climates). Political and economic stability, lower risk for natural disasters and lower risk for terrorist attacks when looking for places to house data.

We are about to go live with Cartika's cloud solution.

oh, and welcome aboard :)

In the future (actually the future is now I think) - people are beginning to value more efficient/eco friendly data centers (which typically require cooler climates). Political and economic stability, lower risk for natural disasters and lower risk for terrorist attacks when looking for places to house data.

Sure, but you are missing the point. The datacenter is very low in the value chain, and it continues its downward slide. We're looking at datacenters being utility companies and I feel sorry for those nations (such as iceland) that will have to put up with these massive energy suckers.. :) Datacenters can be pulled off/onto the grid with little or no effect, their redundancy is meaningless. amzn is scaling up to 30+ goliath datacenters in the coming months

If you are trying to lower the risk of political uncertainty, terrorist attacks and natural disasters - its about the software/platform and not the infrastructure.

Sure, but you are missing the point.

Hi Yash - not really - you are just making a different point. The conversation was about potential future location density for datacenters. Your point is more geared towards the transformation of data centers towards "utility companies"

The datacenter is very low in the value chain, and it continues its downward slide. We're looking at datacenters being utility companies and I feel sorry for those nations (such as iceland) that will have to put up with these massive energy suckers.. :)

why feel sorry for countries or companies that will provide utility services. I certainly do not feel "sorry" for producers of electricity for example. You are correct - data centers are energy suckers - expensive to build and to maintain - hence, the increasing requirement to build and run them more efficiently. You cannot commoditize the service properly until you streamline it as much as possible.

Datacenters can be pulled off/onto the grid with little or no effect, their redundancy is meaningless.

If you are trying to lower the risk of political uncertainty, terrorist attacks and natural disasters - its about the software/platform and not the infrastructure.

now I think you are missing the point. firstly, its not just about redundancy, its about value of the data - which is forever increasing. its about integrity of the data. having it redundant just means you need to be properly secured, insured, certified, etc in multiple locations. Political uncertainty, terrorism, etc, etc are CRUCIAL when considering where data is housed. Secondly, its not about software/platform VS infrastructure - quite clearly it is a symbiotic relationship. software/platforms run on infrastructure and infrastructure houses data in facilities (ie data centers) - you cannot simply separate software from infrastructure unfortunately - and when dealing with data, you cannot simply focus on redundancy being the end all and be all. Sure its wonderful to have a platform that enables global redundancy across multiple data centers - but, this does not suddenly mean that efficiency of data center operations is meaningless, nor does it mean that you have zero regard for where that data is stored. I know I wouldnt want my personal data (be it health data or credit card data or whatever) stored in a 2nd rate facility in a politically unstable country - simply because the platform has the data redundant across multiple facilities... I also would not want to be an end user to a service utilizing inefficient data centers, as my cost would presumably increase because of those inefficiencies correct?

firstly, its not just about redundancy, its about value of the data - which is forever increasing.


Actually, I think quite the opposite. The value of data is forever decreasing with all the junk that's posted online :) But sure, there is important data out there


its about integrity of the data. having it redundant just means you need to be properly secured, insured, certified, etc in multiple locations. Political uncertainty, terrorism, etc, etc are CRUCIAL when considering where data is housed.


Why does security need to be at the infrastucture-level? If the data is important, encrypt-it. If your datacenter is in Yemen and your servers get ransacked by a bunch of pirates, they will find it tremendously hard to get into your data with today's encryption standards (let's talk 1024 bit RSA).


Secondly, its not about software/platform VS infrastructure - quite clearly it is a symbiotic relationship.


That symbiotic relationship was broken some years ago by the likes of Amazon, Google, Facebook.


Sure its wonderful to have a platform that enables global redundancy across multiple data centers - but, this does not suddenly mean that efficiency of data center operations is meaningless, nor does it mean that you have zero regard for where that data is stored.


Sure - but that's because the datacenter is becoming a commodity, the infrastructure is a commodity and cut-throat competitive - its not about SANs or 2*N redundancy anymore.. Most tech companies have learnt that its alot easier/cheaper to build redundancy/security into software, rather than hardware. I personally have zero regard for where my data is stored, as long as I can trust the company that is writing the software.

If efficiency REALLY mattered to a SaaS company, we'd be writing web applications in C/C++. I bet you can cut data-center/infra costs by an order of magnitude if software was written efficiently. But the truth is it's not, because writing software is expensive - and companies optimize for engineers rather than infra costs.


I know I wouldnt want my personal data (be it health data or credit card data or whatever) stored in a 2nd rate facility in a politically unstable country - simply because the platform has the data redundant across multiple facilities...


If your data is important, like healthcare records, personal data and credit card information - it will be encrypted. Hence, secure - regardless where it is located.


btw, I think I've debated you on WHT for several years now :)

Yash-JH;

You are living under a digital rock.

The junk posted to the World Wide Web has nothing to do with the huge amounts of data that is stored on digital media today. The movement toward utility computing is as much about shifting responsibility as it is about capacity.

The state of the computer security art is still in the onion stage (multiple layers, hope they get sick of it) and very few data centers have any defense against explosives and automatic weapons. Encrypting data is worthless if control of the equipment is compromised, as soon as control is lost the data must be considered compromised. Not to mention that loss of access to data can be as crippling as the exposure of data.

Your insertion of Amazon, Google and Facebook into this degrades your credibility terribly.

Your lack of regard for where your data is stored is laughable. You nor I have any idea where all of the data pertaining to us is stored, of course you do not care, you cannot. There is no way you can have any effect on or knowledge of the software either.

C/C++ or Visual Basic - You and all of us are paying for the education systems that are responsible for our digital futures. You really need to look at some of the definitions of efficiency available today.

Debate class may be a worthwhile investment...

Yash-JH;
You are living under a digital rock.


I think so too :cool:


Your insertion of Amazon, Google and Facebook into this degrades your credibility terribly.


Unfortunately, one of those 3 employs me so I guess it does ;). This remains my personal opinion however.

I think so too :cool:

Unfortunately, one of those 3 employs me so I guess it does ;). This remains my personal opinion however.

Considering that Google MapReduce/GFS is rack aware, as is Cassandra, and even AWS developer guidelines make it very clear that people should understand AZs and Region availability issues, it's kinda hard to believe that the application is more detached from the infrastructure and topology than it ever was.

Transit cost and latency still exist, whether we like to admit it or not.

Also, while I agree that it's the software's job to monitor and deal with cluster failures, there's really a limit to our efforts to code around them. We can deal reasonably well with isolated network/node dropouts but large scale cascading failures or heaven forbid fluttering will tax even the best system. If you can't make quorum for any service, you're basically dead in the water.

Latency matters a lot of course. So I guess Yemen probably wouldn't work :)

C/C++ or Visual Basic - You and all of us are paying for the education systems that are responsible for our digital futures. You really need to look at some of the definitions of efficiency available today.


I have no idea what you just said. But if you look at interpreted languages such as Ruby (btw, twitter was originally written in that), PHP etc they make major trade-offs in favor of reduced complexity and rapid application development.

The efficiency of web applications today is horrible, we throw tons of hardware at these systems because its more cost effective than hiring programmers. Also, if you look at how a lot of web servers/apps handle connections - they spawn threads for each request that are put them to sleep when waiting on network/IO resources (one reason NodeJS is catching fire). This is incredibly inefficient, but makes life so much easier.

If efficiency really mattered, gains at the software-level would have a much more dramatic effect.

I have no idea what you just said. But if you look at interpreted languages such as Ruby (btw, twitter was originally written in that), PHP etc they make major trade-offs in favor of reduced complexity and rapid application development.

The efficiency of web applications today is horrible, we throw tons of hardware at these systems because its more cost effective than hiring programmers. Also, if you look at how a lot of web servers/apps handle connections - they spawn threads for each request that are put them to sleep when waiting on network/IO resources (one reason NodeJS is catching fire). This is incredibly inefficient, but makes life so much easier.

If efficiency really mattered, gains at the software-level would have a much more dramatic effect.Sorry for your lack of understanding. It was just a simple statement about your assertion that software written in C/C++ would be more efficient. To gain any understanding of this you need to look to the educational systems that are training current and future programmers to find the definitions of efficiency being taught.

You admit here that hardware usage is more cost effective than hiring programmers, that current practices make life so much easier, yet you seem to have a problem with it. Why is this?

Certainly you cannot mean that things like Twitter, Facebook, Google, etc... have any bearing on the subject of data / data center security.

The Patriot Act has real impact on the security of data stored in or passing through the US. The data contained in tweets or Aunt Tillie's blog is not what is being considered though.

The efficiency of web applications today is horrible, we throw tons of hardware at these systems because its more cost effective than hiring programmers. Also, if you look at how a lot of web servers/apps handle connections - they spawn threads for each request that are put them to sleep when waiting on network/IO resources (one reason NodeJS is catching fire). This is incredibly inefficient, but makes life so much easier.

If efficiency really mattered, gains at the software-level would have a much more dramatic effect.

For what its worth, the list of stable enterprise grade non-blocking IO servers and services is pretty long and old:

Nginx
Tomcat 6 with the NIO connector
Erlang/OTP
NodeJS
IIS 5+ using IHttpAsyncHandler

Apache server has hardly ever been held up as a poster child for efficiency, and really is the only production ready web server I can think of that is still thread-blocking.

Anyone seriously dealing with throughput problems already hire architects that know about async IO.

Nginx
Tomcat 6 with the NIO connector
Erlang/OTP
NodeJS
IIS 5+ using IHttpAsyncHandler


Sure, but I'm sure you'd agree with me that a vast majority of web apps don't use non-blocking HTTP request handlers. In addition, handling HTTP requests async helps, but what about database and IO requests (where a vast majority of app time is spent)?

LinkedIn recently moved their stack from RoR to NodeJS, resulting in massive gains -
http://venturebeat.com/2011/08/16/linkedin-node/

...I feel sorry for those nations (such as iceland) that will have to put up with these massive energy suckers.. :)

Haha, with apologies for joining the party a bit late, don't feel too sorry for us, we have to find something to do with the masses of cheap, green energy we have. If we don't build a pile of Data Centres we'll just end up building more aluminium smelters !!!!

Seems like this thread is straying a bit from the original topic?

In terms of IaaS providers with data centres in Canada, there seems to be a whole lot more data centre providers and ISPs now jumping on the bandwagon, or trying to catch up. Being in Vancouver, Canada, IBM's even put radio ads here talking about their cloud services.

<<snipped>>