Software

10 applications you can move to the cloud

Providers have made great strides in improving the security and reliability of cloud services -- so much so that Justin James sees a number of areas where moving applications to the cloud makes a lot of sense.

Until recently, I was not a big fan of putting mission-critical applications in the cloud or letting someone else provide them. I had been burned too many times by shady vendors or providers who just did not have their acts together. But in the last few years, things have changed. There is a new breed of application vendors out there who have certain application classes nailed down really well and have established reputations for reliability, security, and fairness. It's a good time to take a look at the cloud again. Here are 10 applications that can be moved to the cloud.

Note: This article is also available as a PDF download.

1: Email

Email is the lifeblood of many organizations, and as a result, many companies are not willing to let go of it. That is understandable. But hosted email providers have moved beyond the days of packing 5,000 mailboxes belonging to 300 accounts onto a cheap computer running with a basic POP3/SMTP setup. While basic email service is still out there, you can get hosted Exchange services from a variety of vendors (if you need it), as well as some upscale, non-Exchange offerings. Email architecture has become quite standardized, and there is really no value-add to keeping it inside your firewall other than mitigating regulatory concerns.

2: Conferencing software

Setting up and maintaining conferencing software is not fun. To make matters worse, when it is down, it needs to be up in a hurry. Like email, there is zero benefit to locating this within your firewall. Also like email, the setup and configuration can be complex enough to require an expert, unless you don't mind tying up a staff member for a few days. For a low monthly or yearly fee, this weight can be off your shoulders. No one will notice or mind, and your staff can move on to other tasks.

3: CRM

The decision to outsource CRM can be scary. After all, like email, CRM is where so many of the company's crown jewels are stored. But there are no technical benefits or advantages to having CRM in-house. It's a fairly low bandwidth application with maintenance overhead you do not need. In addition, the licensing of many CRM systems can be a hassle. Moving to a hosted CRM system can free you to spend more time on more important issues.

4: Web hosting

Hosted Web space used to be as awful as hosted email, unless you were willing to spend big bucks on a dedicated server. Many vendors have shifted to (or offer) a virtualized hosting environment, which has dramatically increased uptime, reduced security risks, and allowed them to provide much more open and direct access to the servers. This is great news, especially for companies with custom applications that require a deployment path beyond copying some files over.

5: Development test labs

Building and maintaining test environments for software developers is a major undertaking. To do it right, you need all sorts of permutations of operating systems, patches, and relevant applications. You could easily find yourself with nearly 100 test beds for a simple Web application, for example. Why do this to yourself when there are quality vendors out there who already have these test systems set up or that allow you to configure them with point-and-click ease? And you can safely give the keys to the development staff and know that they can't permanently mangle the test systems, too.

6: Video hosting

A few years ago, I was down on the idea of using the common video sites to host your video. Many companies would block those sites under the assumption that they were only for games, there was the real fear of having ads placed on your videos, and often the quality would be compromised. Now, the big name sites have upgraded their quality and few companies block them because there is plenty of legitimate usage. In addition, some sites allow you to pay a fairly low charge to give you more control over your video, like deciding where it can appear and enhancing its quality.

7: Email security

Even if you do not put your email with a hosted vendor, you will want to look at having a third party perform your anti-spam and antivirus duties, even if it's only as a first-line defense. If you look at how much incoming email is spam, you'll see that you can reduce your bandwidth needs dramatically by allowing a third party to perform an initial scan of your email. It will also allow you to have far fewer email servers. Speaking from personal experience, even a small company can have its email servers and network overwhelmed by incoming spam. Getting a good spam scanner outside the network can make a night-and-day difference.

8: Common application components

There is always the perpetual "build" vs. "buy" question for development projects, but the cloud adds a new wrinkle. Many functions that used to be the purview of components or libraries you could buy are now being made available as Web services, typically billed on a per-usage basis. Often, these services take a variety of lower-level functions and roll them into a complete, discrete offering. You would be surprised at how many of these Web services are available, and depending upon your usage scenario, it could make a good deal of sense to use them instead of rolling your own.

9: Basic office applications

If you need the full power of the Microsoft Office suite, by all means, this isn't for you. But if you are one of the many organizations that use only a small fraction of the Office feature set, it may make sense to look at one of the new crop of online Office replacements (or even Microsoft's online version of Office). I honestly never thought the day would come when this was possible, but it is a legitimate possibility for some companies. Just be honest with yourself before making this move and work closely with your users, since this directly affects so much of their workday.

10: Batch processing applications

One type of application that shines in the cloud are batch processing applications, such as data warehouses. As long as you can get the data needed into the cloud without disrupting your operations (such as "seeding" it with the shipment of physical media or synchronization over time), the ability to rapidly scale capacity in the cloud can result in tremendous savings. For example, if you need 15 servers' worth of computing capacity for a once-per-week job, do you really want to have 15 servers sitting idle in your server room waiting for that? Or would you rather just modify the task to start 15 cloud instances, run the process, and shut them down again? In a scenario like this, it is clear that cloud computing can deliver significant advantages.


About

Justin James is the Lead Architect for Conigent.

57 comments
joshua_keefer
joshua_keefer

One word. Security. It will be a few years and few incidents before some REAL security for cloud apps comes around. I say keep as much as you can behind your firewall in your physical control until something better happens.

vzjp
vzjp

How about ITIL based Service desk management? An example is over at VanillaDesk.com

BrightLibra@Gmail.com
BrightLibra@Gmail.com

Would one of those 10 be financial applications and databases? Could you see putting enterprise financial data, customer base and even corporate intelligence in "the cloud?"

scottchapm
scottchapm

Any examples of where to look for this sort of thing? Sounds great but its not something I have ever come accross.

techrepublic
techrepublic

@Justin James: Your vendors providing "reliability, security, and fairness" could evaporate on a whim or a bad turn of the economy. Then, if you're too deeply invested in their solutions, you will be left high and dry. The so-called "cloud" (it's a NETWORK, fercryinoutloud!) is notoriously unreliable and if you depend on it for application-critical uses, you'll *eventually* be burned. Consider this fair warning.

iwan_budihalim
iwan_budihalim

I believe someday in the future, we have to move our core application into cloud. It's a matter about cost reduction. It's the same thing with Mail Services, a decade ago, every company didn't believe in Mail Hosting, and try to take care of them self. Now is different, now we're ready to move our mail system to cloud.

pjboyles
pjboyles

The only item I see for an external cloud is #2 Conferencing. All of the rest should be internal. There are legal, backup, support, interconnect, SLA, management and other items yet to be answered for external clouds. Let's not forget the biggee, the cloud provider does not own the customer's data and the customer's data will be immediately made available to the customer on demand even if the cloud provider shuts down. Losing access to your data is a big oops. More questions for the cloud vendors to anser: How does mergers and acquisition fit in? Do all contracts and terms survive? Bankruptcies? International considerations? So many things need to be put in place to make external clouds viable for most uses. Also, External clouds should offer a standard set of global APIs, data storage, data management, backup/restore, migration, change management and transparent problem management. When that happens an application is written and the data is designed to sit on top of any provider's cloud and it all becomes more attractive. Current provider raise costs too high? Bang you migrate to someone who fits your budget immediately without a huge cost and the time to port everything. Another issue is that many business bring all WAN links back to a central location then out to the internet (central firewall and proxies). Everything that is then external adds yet that much more bandwidth requirement. And external clouds add yet more locations which must be accessible and up for people to work. From the perspective where I work, each location can work for several hours without WAN connectivity. Are they able to do everything? No. But product still gets made and delivered to the customers.

edwardwstanley
edwardwstanley

Clouds are ok. until it "rains" or "lightning strikes" anywhere between you and your private information and business it processes. Internalized solutions lets you cope with just your local weather.

bfunke
bfunke

Next is another infommercial for Google apps that the Tech Republic likes to run. It's been at least a week.

markb
markb

If I spend money each year on security to prevent data leaving my systems, why would I want to give it away to an anoymous cloud service?

evanmathias
evanmathias

So for email a business can buy a hosted solution which interfaces with their active directory and MS Outlook, providing archiving (eg vault), backup and recovery. All this through the familiar MS Outlook, no logon required (since AD), and able to dictate their email address? Scalable both up and out? Or are we talking about a Hotmail scenario?

Integrateful
Integrateful

Two or three examples in each category would have made the article more useful ... even if they were just the authors own preferences.

jpoling
jpoling

What gets run in-house then if an organization moves all 10 items to the cloud?

luke
luke

The government would love you to do all that. Lot easier to get to "your stuff". At least with stuff left on premise, they have to bust your door down to get to it (if you've done your homework on the firewall side that is). Email is already being monitored. So that doesn't really matter where it is. Some of those "cloud" examples therefore have merit - spam and av filtering for instance. Web Hosting obviously. But the rest? CRM? Databases with company info on em? Sure. If you can control and monitor all traffic to the virtual machine and make sure that copies of that VM aren't getting into the wrong hands. Not likely.

blarman
blarman

1. Email. Absolutely. 2. Web-Conferencing. With GoToMeeting, WebEx, etc., this also can significantly decrease your IT workload. Be careful of new OS, however. It took vendors some time to patch/build for Windows Vista and 7. 3. CRM. Maybe. If it is self-contained, move it. If it integrates with other parts of system like payroll or AR, not such a good idea. 4. Web-hosting. Yes. 5. Test labs. No for two reasons: one is sensitive data, two is replicating production. You want the test lab to mirror production as closely as possible. Unless production is in the cloud, test should stay local. 6. Video hosting. This one -may- depend on content. Generally, I agree. 7. E-mail security. Goes with #1. Do it. 8. Not sold on this one, mainly out of performance concerns. 9. Only for those who don't travel. Web apps aren't available in most airports or airplanes and air-cards for laptops are flaky where I live. No to this one. 10. This one is going to be very specific to the task. It's going to make sense in some situations.

benchsoft
benchsoft

As a developer of accounting software, we are faced with new competition from cloud accounts packages. Opinion varies wildly about the future of this market. Any from thoughts or experience?

mswanson
mswanson

What about hosted PBX? With business continuity a big concern, plus mobility, this seems like a no brainer.

geldernick
geldernick

All sounds good ... except: there is never a problem until there is a problem. So..oooo. I recommend mirroring the site ... then instead of increasing business risk and survival you add a level of confidence that your organization can retrieve your critical data. Donald Geldernick Senior Program Manager geldernick@aol.com

terjeb
terjeb

In some cases you are absolutely right, these applications could (and some of them should) be moved into the cloud, but there are obstacles, for some these are not easy to overcome. 3/ CRM (since that is what I am doing these days) Sure, but there are a lot of "ifs" here. In many cases the CRM is introduced into a mix of a large number of in-house, perhaps industry-specific applications. In many cases these apps "own" the customer relationship (billing etc). Now you have to deal with integrating the CRM with these apps, and that will often be (a lot) harder when the app is in the cloud. In many cases your organization is not even the owner of the data, so there may be political obstacles to moving the data into the cloud. Also, some organizations are allowed to store sensitive information in-house, such as social security numbers. These may be required in some client communication where your CRM would be your primary application, and storing such information in the cloud, or even passing it between your organization and the cloud regularly, is often not an option. Which brings us to 1/ Email Yes, unless your CRM is in-house, in which case integration becomes an issue. Email integration with the CRM is vital. 8/ Common components YES! This is an excellent idea, with the caveats mentioned above. 9/ Office applications Well, perhaps, and perhaps not. Again, sensitive data and also integration becomes an issue. Your points are really good though, and with the caveats mentioned, I totally agree. Dumb as I am I had in fact never thought about 8, which I am now going to push hard here. There is a of stuff of the "common component" nature we could/should move into the cloud.

rograham1
rograham1

It would make sense to consider the cost of the tiered software licenses/support in any decision to relocate workload to the cloud. Your mileage will vary and this consideration could sway your strategy in either direction.

DonSMau
DonSMau

Well, I suppose they / could / easily move to the cloud. But moving any business-critical, or business-model-defining, process or data from in-house will eventually end in grief.

Tony Hopkinson
Tony Hopkinson

Given the security/data protection issues are addressed, it's a very elegant solution especially with varying demand and quite often a perpetual increase in the volume of data. Generally systems like this aren't critical to the operation and given the way they are generally implemented, dynamic DBMS replication would be a very painful exception, it should be relatively risk free. So get the benefit of the cloud, use it for what it's really good at, not just move things into because you can and then cross your fingers and hope for promotion before the wheels come off.

QAonCall
QAonCall

"Building and maintaining test environments for software developers is a major undertaking." Definitely a truism..however, most are also a complex collection of interconnected systems, which adds to the complexity and the challenge. "To do it right, you need all sorts of permutations of operating systems, patches, and relevant applications." Again, yes, with some caveats. Your inherent assumption is that the application built is 1) Web facing 2) Not 'protected' (public facing) and that 3) You are trying to satisfy an uncontrolled market. This is in fact, not the case. Many companies only code to 1 standard, and expect their customers to comply (I know people are going to go nuts over this, but my last three clients all dictated this). "You could easily find yourself with nearly 100 test beds for a simple Web application, for example. Why do this to yourself when there are quality vendors out there who already have these test systems set up or that allow you to configure them with point-and-click ease? And you can safely give the keys to the development staff and know that they can?t permanently mangle the test systems, too" Ok, here we may be in agreement, especially if you are a small company, with limited resources and the application you are developing requires little or no interaction with local data, and performance testing is not part of your scope of work. Further, if you are a larger company those would all be factors, as well as the use of automated solutions to push builds, perform smoke tests and regression as well as the ability to do performance and load testing. I see as a more common solution the use of virtual machines and have done extensive work in this area. This allows the push of builds to certain selected machines on my schedule. These assets can also comply with some legal issues around SOX and other regulatory stuff. Building genreic machines is simple, and generally one time. Resource needs are very small. They are flexible and on-demand. They allow for checking reported issues on demand. They are completely customizable, easy and cheap to maintain and store, and allow fluid and secure interaction by your teams. Cleanup is simple and easy. All the benefits of your example, but none of the risks, or potential risks. With my background, I can see an opportunity for some clients in the cloud, but I think this is a thorny area, without listing some pretty important caveats. Overall, great list though! kudos!

daboochmeister
daboochmeister

I can think of a few, but can someone point at (or make here) a list?

Inventory Software
Inventory Software

As big shakes are made by web shops changing the shape of retail, these etailers are having to sort out their backend warehouse operations. They have an online financal system - great. Next, migrate away from painful client server model for their warehouse management software and get in the clouds. Seamless integration via open APIs and Web services...music to web shop ears. The real winners are the hybrid approach so if the internet connection is down, the operation carries on regardless. WMS, from companies like PeopleVox, is the next acronym to go full steam into the sky.

tbmay
tbmay

Such as e-mail (smtp is insecure anyway) and web hosting. (What's the point in not doing it?) But other parts of it I'm very skeptical about. The businesses I deal with have a lot of private, sensitive information stored in MS-Office files, and I don't think it's good for them or me to not be able to work because their WAN link is down, for whatever reason. Same principle applies to CRM.

Justin James
Justin James

... but ask yourself, "what security concerns are mitigated by handling this myself?" Honestly ask yourself that question. Chances are, the answer you will get are "none", so long as the cloud vendor is doing a few things: * Isolating your data (different DB instance, or DB in an isolated VM) * Isolating your settings (separate VM instance for each customer will work fine) * Encrypting traffic on the wire * Running the same level of background checks and other due diligence on their employees that you would Of course, these are all different from vendor to vendor, but if the vendor *is* doing these things, chances are that it will actually be *more* secure than any deployment you do, because they will have more specialization in that application. How many times has a non-expert in-house done something like open the firewall up because they aren't quite sure which ports are really needed by a complex application, for example? J.Ja

erik
erik

If you are in a metro area perhaps the WAN is "Insanely Stable". However there are lots of places where this is simply not the case. Just today the Qwest backbone feeding our area was down for 6 hours. If we had everything mentioned in the post in the cloud we would have had to send most of our office staff home, or just tell them to sit on their hands waiting for the internet to come back up. As it was it did break a tunnel to a remote location, but locally our internal database applications & email were unaffected so people could perform the greatest part of their normal day to day work. Not everyone is in the e-commerce business & can claim that business would come to a halt anyway....

sars42
sars42

We have a few services in the cloud but I'm not a fan of putting everything into the cloud. It's true that relaible comms links to cloud services are critical but it is also that same potential for an outage that makes cloud computing attractive. If everything was in the cloud, disaster recovery would be so much easier.

Justin James
Justin James

There are a number of providers doing this already. I'm not 100% sure about the AD authentication piece, but outside of that, it's there. J.Ja

c2moore
c2moore

Has anyone got more examples? Gotomeeting, webex,office, google apps, .......

tbrown
tbrown

Not to sound like a complete ignoramus, but what are the issues with CRM systems in the cloud? My main issue is in terms of not being able to integrate data seamlessly with onsite resources without jumping through hoops. Wouldn't CRM privacy concerns be handled merely via not logging sensitive emails through the CRM system or just setting it to register that a message was sent, and not the message text?

mswanson
mswanson

Hosted PBX is ideal for the cloud.

BlueCollarCritic
BlueCollarCritic

The bottom line with this issue of privacy specifically and other areas that involve government action like those of protection of freedoms & liberties is that less is more. As great as our form of government was initially, its natural path of progression is to provide the citizenry not with more but with less privacy while at the same time providing itself with less transparency. In other words, government involvement produces the opposite of the desired effect. This happens not because government is necessarily evil but because it like everything else in our world operates on the growth principal. That principal states that Growth = Success In our world we (the humans on this planet) have allowed the rule of growth to dictate how success is achieved or measured. If you want to succeed be it as a human (grow your family) or a business owner (grow your bottom line or market share or both) or an entertainer (grow your record or video or ticket sales) or philosopher (grow your charities donations) or even a religious leader (grow your church?s membership) the bottom line is the measurement of growth. And to that end we?ve allowed growth of government to be a measure of success where it used to be viewed as a measure of failure. In the example of business it is possible to be profitable and do so year after year but not technically be successful for you did not grow your bottom line or market share. This is the cruel reality of how most business philosophies are taught. The idea that if you grow then you will be profitable by default; at least most of the time. The founding fathers saw big government as bad where as many today no matter what their political preference see non-small government as being anywhere from necessary to good and so some form of growth in our government from its inception has been (until very recent) viewed by most as not only necessary but a good thing. Now how much growth is good does vary amongst the population but I have no doubt that more then 4/5ths would say as of today that the government should do more now than it was setup to do in the very beginning. And so once we established growth was a sign of success the natural evolution of survival took over and so each generation of politicians sought to grow the government to show their own success as a leader and to ensure their political survival. At some point in the late 1800?s to the early 1900?s this growth snowballed thanks to the financial help of some very powerful and intelligent persons who saw what was happening and realized the growth could be manipulated to their own financial growth or success. Jump forward a few decades to now and we see that each step the government takes to supposedly re-enforce our liberties or rights often has the opposite effect (or at least no real effect) and so we are left of the same or more often worse than we were before. And so I say again that in the area of privacy less is more especially when it involves the use of government.

BlueCollarCritic
BlueCollarCritic

The cloud is but one of many tools/methods/devices (pick your favorite term) that companies will use to do business. It is NOT the end all be all that many try to sell it as. Your WAN may (and I emphasize MAY) work fine now for what you are testing but currently only a few are trying to use the Cloud and even then only for parts of their services. When even a sizeable number try to move to the cloud you'll see just how susceptible those stable WAN connections are and then the Cloud push will take a great big step back; most likely to a place where it should be and not where Cloud Sales People want to push us too. The Cloud just like the following list of IT fads will in the end turn out to be useful but not to the extent it was sold as in the beginning. Remember any of these so called ?Game Changers? in IT? 1)Dummy Terminals ? Someday (soon) no one will use personal PCs and everything will be either a server or a dummy terminal. 2) .Net ? The .Net platform form Microsoft will change the software world. 3) XML ? Soon XML will make all other forms of data storage and communications obsolete 4) Virtualization ? VM will make desktop PCs a thing of the past 5) The Cloud ? Cloud Services will make the need for local servers and PCs ancient history 6) ???? - I wonder what the next promise of IT Heaven will be

dhearne
dhearne

...you go from owning a process or application to being a customer who uses it. When something is broken and the CEO is pissed, telling him you're on hold with your cloud provider is not nearly as comforting as telling him the entire IT staff is working on resolving the problem. Have a close look at your SLAs. Response times are not the same as 'fix' times. If everything is in the cloud, all you can do is sit on your hands while someone else fixes it at their whim. And, if that is the case, why does the organization really need you anyway?

Justin James
Justin James

I agree completely. At the same time, I've come to realize two things: 1. WAN links are insanely stable now. Our office only has WAN downtime when we take a critical network device down. We've had under 10 minutes of vendor created downtime in the last 2 years or so. 2. The WAN link is so vital to many, if not most businesses, that they are dead in the water when it is down anyways. That was something that really turned around my attitude on cloud, the simple realization that a dead WAN = dead company, regardless of what apps are in it. Heck, half your applications can't even bring up a "help" screen now without the Internet... J.Ja

evanmathias
evanmathias

Sychronised AD would make this cloud solution identical to most existing solutions, but at a fraction of the cost. Allowing businesses to offload Data-Centre costs and easily provide a single email domain address spanning locations. Ok had a quick look around the internet. I know Outlook 2010 intergrates with Exchange. It seems there are suppliers who syhcnronise with AD and provide the local experience via the cloud. Does anyone have any experience with an Outlook 2010 hosted solution? It does seem to have the potential to be a no-brainer choice as a company email system.

luke
luke

CRM systems start as simple contact databases and turn into "let's keep everything about the client" depositiries. Good luck trying to seperate your sensitive data from non-sensitive. Where would you keep that sensitive stuff? On another CRM?

BroadcastArashi
BroadcastArashi

I wouldn't think a PBX would be suitable to a cloud. Correct me if I'm wrong here, but wouldn't that blow about 150ms of your round trip latency just getting to the cloud? That's also if it's within the country.

kennedy.barry
kennedy.barry

I agree with your statements regarding IT fads... somewhat valid examples, all. However - is the trend to shift resources to cloud providers actually a fad? Was outsourcing IT needs to India a fad? It's another example of outsourcing and commoditizing IT costs. You are making the same type of extreme and overly simplistic statements as was made by the very people who prophesied the fads you referenced, when you tie the success of cloud computing to the speed of your WAN connection. As if any company needs to be limited to having just one WAN provider. Why would anyone in their right mind have one single point of failure for something as critical as connectivity? What? We're going to go back to writing letters on vellum with a quill, when our WAN connection is overloaded or our provider has an outage? It's relatively simple to project bandwidth requirements. There is a recent development called math, you might have heard of it.

Kim SJ
Kim SJ

I was horrified by your post. If you gave those arguments to me, I'd fire you at once. Keeping the CEO happy is not the same as doing the best for the company. IT is a core function; get with the program and start enabling the company to be best in class.

tswartz
tswartz

The organization won't need you in the near future anyway. Also, just because you see someone fixing it, doesn't mean its fixed any faster.

tbmay
tbmay

Call me old school for preferring to not have all my eggs in one basket though. There are a few 2010 trendy things I don't consider great ideas for businesses that take their data seriously. 1. Running the perimeter security as a vm on a host with other things. 2. Trusting online backup services. 3. Moving everything to the cloud. Security is obviously the biggest problem with all these things, but it's not the only one. Security by itself would be enough for me not to like them though.

bigaussie
bigaussie

We (via our partner NTT/Verio) provide Exchange 2010 globally. I have been using Exchange with many of our hosting clients since Exchange 2007. The system works just like the Exchange server was located locally. Email boxes up to 25GB and the ability to easily connect Blackberry, Android and iPhone. All with a monthly charge which starts at around $10 per user account. All the things you would expect from a local Exchange server - Calendar collaboration, Web Access, ability to have it all permanently archived (as per many Government regulations), first level defense for Spam and Anti-Virus built in. There are plenty of other companies out there providing this service. Just check the whole range of features and how much some of the features cost to "add-on". My clients are very happy not having to worry about employ an Exchange specialist. In the past 3 years of using the service it has not gone down once. I have had a couple of dramas with clients having Outlook problems "locally", but because the Cloud was still working connected to their Blackberry they could continue working while their IT people fixed the local copy of Outlook. Nothing but smiles here :p

tbrown
tbrown

Didn't consider that perspective

Neon Samurai
Neon Samurai

My quick/dirty would be a *nix box doing routing with two NICs. This would require the ISDN and Cable provider's full setups on my side of the wall. An ISDN modem and it's hosted subnet (if it won't bridge) plus a Cable modem and it's hosted subnet (Roger's won't bridge so I have NAT inside NAT now). I thought this a nicer solution for an office since it places the ISDN and Cable modems insde the same gateway device. Slap the router on the wall where your cable and phone come through the wall and your good. Now, if you just wanted to double your pipe then buy two feeds from the same ISP and run them through a *nix homebuilt or any other coupler. If the ISP goes down then your down but when they are up, you should have twice the bandwidth. If it's for redundancy then you don't want two feeds from the same ISP. You want one feed from two separate ISP and ideally on separate major branches. The down side is that you can't bind both feeds to double your bandwidth but when one goes down, you have the other working still (hopefully).

Tony Hopkinson
Tony Hopkinson

dumb ass techy types to take the fall if required. I'm not just a nice hairdo you know. Which clueless numpty wrote this software by the way. My reply doesn't seem to be aligned with my requirement. You propeller heads need to get your heads right, us business brains can't keep carrying you forever! Is this obvious enough for you, or should I write IRONY on the back of a paper plate with a custard pie on it, jam it down your shorts and pull them over your head? FFS.

kennedy.barry
kennedy.barry

I was working as a Network Analyst at an airline when a local Telcom cut both of the fiber lines, which were not supposed to be run in the same trench. The airline went back to writing tickets by hand, canceled flights, and sued the Telcom that ran both lines in the same trench. Had it been an internal employee that had decided to follow your logic, he would have been fired on the spot... and probably been blackballed.

kennedy.barry
kennedy.barry

one that supports fail-over and load balancing. If it's a smaller company, you can get a used Firebox, or take an old computer, throw a couple NICs in it and put Linux on it.

tbmay
tbmay

..for anyone betting the farm on the cloud.

Neon Samurai
Neon Samurai

It's pretty easy to do redundant setup. I just saw a router with dsn and coax headers on it. One plugs into Rogers, the other plugs into Bell. If one goes down, the router has the other ready to go. (now, if one could couple the two destinctly different ISP/subnets into a single connection when both where working.. mowahahahaha) Granted, this is for a medical facility that has to have connectivity. I'm not sure what costs other than paying two separate ISP subscriptions would be incurred after setup. Sadly, in my local area the power grid is far less stable than the phone or cable grids. We've had daily power failures around ten with an additional outage yesterday just before dinner time. Anyone know how to go about invoicing the local electricity provider for the expense of having to buy a UPS? If they can't bother to fix the local transformer maybe they can pay for the UPS that will minimize the damage they cause.

Tony Hopkinson
Tony Hopkinson

Because it's cheaper, and technology is so much better now preparing for a failure is not worth the cost..... Never going to happen before you get promoted for saving all that money....

four49
four49

"Keeping the CEO happy is not the same as doing the best for the company." Depends on the company. Many large bureaucratic organizations are all about appearances, with little if any focus on results (think phone companies here). Sometimes the CEO is MUCH more interested in seeing something getting fixed rather than it actually being fixed. People who get promoted are those that appear to be important, as opposed to the ones that really keep the place running. Luckily, companies like this are a dying breed but they do generally take a long long time to die.

Justin James
Justin James

... we're old-school in the exact same way! "1. Running the perimeter security as a vm on a host with other things." I am in the middle of replacing a physical firewall with a VM'ed firewall. W made the deliberate choice that the *only* VM to ever run on that host is the firewall. We're only going with a VM so we have hardware independence and can rapidly move to another machine in case of failure. "2. Trusting online backup services." My first draft had *offsite, archival* backups as something you could put in the cloud, but said that your current working sets should be there. It's not even a trust thing, it's a SPEED thing. When things are down, the last thing I need is to wait 10 hours for a download! "3. Moving everything to the cloud." Agreed. I question companies that have nothing so "sacred" that they refuse to put it in the cloud regardless of the advantages. If nothing you have is so valuable that you are nervous about someone else handling it, what does your company do that's valuable? ;) J.Ja

Editor's Picks