We want to meet you.

We have a new role currently vacant – read on to see if it might be for you.

Position Description
The Training Manager will report to the General Manager – Volume and would be responsible for:

• Training development and delivery across all uber departments.
• Quality Assurance management to uber ServiceDesk staff.

Duties for the role include:

• Working with team leaders to develop service and technical training packages for staff
• Delivering and coaching to service and technical training with staff
• Provide ongoing quality assurance of staff performance
• Incorporating coaching for staff
• Development and delivery of training packages for new products/services
• Add to and update the KnowledgeBase
• Development and delivery of induction training packages
• Working closely with GM Volume to improve the efficiency of operations

It is important to note that the training manager must be prepared to travel to the United States to provide training to uber contractors and/or staff if required.

Experience
Applicants should have relevant experience in developing and delivering technical training. Familiarity with the following technologies would be highly regarded:

• Domain Hosting
• Email Clients
• Microsoft Business Office Productivity Suite (Exchange, Sharepoint, Lync, Dynamics)
• Linux and Windows Operating Systems
• DNS

Skills
Applicants should have:

• The ability to work as part of a team
• Good written and verbal skills
• Effective time-management skills
• Agility
uber is a rapidly growing company, with our success being built off the ability to quickly recognise and embrace opportunity.

Cert IV in Training and Assessment will be highly regarded.
Relevant technical qualifications are important but will be assessed in conjunction with experience.

If your pulse has risen and you are excited about this position, please email your résumé to k.clark@uber.com.au to make your mark in uber history.

While this change has been ongoing since its initial testing phase, which started at the end of May 2011, we’re now using this fantastic tool to maintain the usability and reliability of all of our cPanel shared servers.

I’d like to take a moment to explain a little more about how Cloud Linux works and why we have chosen to use it.

On shared hosting it’s easy for one customer with peak site traffic or a large site to use more of the server’s resources than those with small websites. Previously, if one user on a server was overusing resources, all customers on that server would experience slow service, and possibly even down-time if the server had to be restarted.

Now, the Cloud Linux system controls the amount of resources an account can use on the server, ensuring that resources are allocated fairly between users. This means that anyone overusing resources will not affect other users’ service performance. Those overusing resources will only affect their own service’s performance, causing it to slow, and eventually to receive and error 500 message when the maximum limit is reached.

For some customers who were previously not limited in the memory they used on the server, the Cloud Linux installation has caused some of the above issues.

As with all great things, it’s taken a little time to perfect the settings on Cloud Linux for best performance. Initially we did some testing on a few shared servers, and found that Cloud Linux was working well as a way to maintain the servers’ stability and quality of service. Since then we have gradually been installing Cloud Linux on the majority of our shared server fleet.

At the start of the original testing the limits were set at 256mb; since then we’ve increased twice, based on our monitoring and feedback on these systems.

Today’s Cloud Linux customers will find themselves being able to take full advantage of a 1 GB memory limit.

If your service is still experiencing poor performance and/or you are getting an error 500 message, please contact support so they can investigate the cause of the problem and advise on further options.

We’ve received feedback from customers who found that our rollout of Cloud Linux created issues for them, due to the changes being made without prior notification.

While our decision to make the changes without notice to the appropriate customers was certainly erroneous, it was made with the success of our small-scale deployment in mind.

We understand that this has been frustrating and time-consuming, and apologise to any of our customers who found this to be the case.

We expected a much smoother and more effective process, with little to no negative impact for our customers.

This problem is something we will be sure to keep in mind when we’re next looking to provide any new service to improve customers’ experience.

We anticipate a huge improvement to customers’ service and relationship with us moving forward.

Just a quick post to let you know some exciting news about our newest addition to the online community for Jumba, AussieHQ and UberGlobal.

April Helene-Horton has joined our team after almost 5 years of experience working with one of Australia’s leading customer service organisations.

She’s committed to making sure your feedback and questions are addressed, that your opinions are heard and that you’re kept up-to-date with every move we’re making to maintain the exceptional experiences you’re used to.

April will be chatting to you on Whirlpool, the Jumba, AussieHQ and UberGlobal forums, Twitter (@jumbahost, @aussiehq, @uberglobal) and Facebook (Jumba, AussieHQ, UberGlobal), and she’ll also be putting up some super snazzy blog posts for your reading pleasure. She’s going to start answering questions today, so please make her welcome!

Our clients get up to some weird and wonderful things, but this month one of our dedicated hosting clients, Paul from Synchromedia, is taking it to a new level.

Today he is embarking on a 10 day bike ride that would make most of us wonder about his sanity! The ride is from Brisbane to Mission Beach, 1650km over 10 days.

Paul is riding for a charity called Tour de Cure which raises money for cancer research. Over the past five years Tour de Cure has funded 61 research, support and prevention projects for men, women and children totaling over $3.8million. You can catch the Tour de Cure action each morning on the channel 7 Morning Show.

I had the good fortune of meeting Paul when he cycled past our office in Canberra on his Tour de Cure campaign last year. His down to Earth demeanour and infectious positivity endeared him to everyone here at Uber, and we wish him all the best for his journey this year.

Uber is 100% behind you Paul, good luck with your fundraising efforts to help find a cure for cancer.

If you would like to donate to Paul please go to his website and read his story at www.mytourdecure.com.au.

Between now and June 30 you can buy 6 months and get 6 months free on selected shared and dedicated hosting packages with AussieHQ and Jumba. All you need to do is sign up for a one year billing cycle and then use the promo code EOFY2012 at the checkout to unlock your discount. Choose from any of the following products:

Jumba

cPanel Personal Hosting
cPanel Extreme Hosting
cPanel Small Business Hosting
cPanel Reseller Hosting

AussieHQ

Business Linux Hosting
Business Windows Hosting
cPanel Extreme Hosting
Dedicated Hosting (VMware VPS, DSE1, DSE2 or DSE3)
VMWare MPS Power Reseller

To receive your free 6 months, simply sign up for any of the above packages on a one year billing cycle, and then enter the promo code EOFY2012 and hit ‘validate code’ to activate your discount.

Our EOFY offer expires 30/06/2012 at midnight AEST, so get in quick and grab yourself a bargain!

This promotion does not include domains, dedicated IPs, SSL certificates or dedicated add-ons.

Cloud computing allows for resources, software and data to be provided as a service rather than as a product to be managed by the consumer. The advantages of this model for businesses include faster application set-up and better resource allocation without the hassle of ongoing maintenance and management.

Telsyte senior analyst Rodney Gedda reports, “With more than 50 per cent of Australian enterprises using server virtualisation technology, the interest in private Clouds is also strong with 19 per cent of organisations building a private Cloud and a further 35 per cent considering a private Cloud”.

Virtualisation is an overarching trend in the IT sector, as it allows computers to achieve more than previously possible, making it a more cost effective option for businesses. Hardware virtualisation allows multiple virtual machines to be created within the one physical server, with the software on each machine being separate from the server’s core hardware resources.

Although IT professionals are abreast of this trend, many business owners who run Microsoft Software do not realise that they can now run Windows on a virtualised server to accommodate their business needs.

Microsoft software dominates the Australian business sector, creating an opportunity for web service providers to offer Microsoft-based virtualisation products to this key market. Australia’s fastest growing web services provider, UberGlobal, is backing the virtualisation trend and brining this business opportunity to Australians with their new Windows Web Server product. The product uses Microsoft’s virtualisation manager (called a hypervisor), Hyper-V, to slice up physical hardware, creating multiple virtual machines to on-sell to business consumers.

“We think there is a massive opportunity for us to bring Windows-based hosting to Australian businesses”, says UberGlobal CEO Michael McGoogan. “Microsoft technology can now do everything an Apache server can, meaning it can run MySQL, PHP, and all the popular web applications such as WordPress, Joomla, ZenCart, phpBB and Drupal. It makes sense for businesses that are already running Microsoft software onsite to leverage Microsoft virtualization software via one of our Australian based Enterprise-grade data-centres. We’re so confident that Australian businesses will love the simplicity of running Microsoft software on Microsoft hardware virtualisation that we’re offering a free one month trial on our new Windows Web Server product”.

UberGlobal is a leader in automation IT services to the SMB/SME market. This year, they were awarded with the Parallels global Cloud Leadership award.

Evan Williams, Small and Medium Business & Partner Group Director, Microsoft Australia says “UberGlobal is a valuable Microsoft provider, with a long track record of success. Its strong presence in Health, Government, SMB and large enterprise sectors, allows them to provide advice and support for a wide range of Microsoft customers,” said Williams. “UberGlobal’s recent achievements with hosting solutions, including Hyper-V, demonstrate the value they provide to their customers, offering savings, productivity and improved collaboration.”

At 2:41pm on April 4^th 2012 the Uber Global monitoring systems detected a large number of network communication failures primarily associated with traffic passing through our Sydney Data Centre. The root cause was a malfunction associated with the operation of HSRP (Hot Standby Router Protocol) on one of Uber’s Sydney Border routers.  The HRSP implementation for a number of virtual interfaces on one of the Sydney border routers stopped passing traffic even though the interfaces themselves did not fail or change state.  The monitoring system immediately detected the drop in traffic flow but no failure in the operation of the router itself was reported.  This led our Engineers to suspect the issue was associated with a network/layer 2 failure rather than a fault with HSRP on one of the routers.  Once the issue with HSRP was determined services were quickly restored.

Background

Hot Standby Router Protocol (HSRP) is a widely implemented Cisco Systems protocol (RFC 2281) designed to support high availability by allowing a single logical interface to be shared between two physical routers.  In a normal operation HSRP  should ensure that routed services are maintained even when a physical router or interface fails.

Incident and Service Restoration Chronology

The timeline of the Incident is as follows:

• 2:41pm – Monitoring system detects several network and service failures.
• 2:42pm – Investigations begin; alerts for core services are checked to determine root cause.
• 3:20pm – Further investigations found network communication issues with several services.
• 3:45pm – Sydney border router HSRP failure was discovered.
• 3:55pm – Sydney border router reloaded, incident resolved.

Root Cause

The root cause was a malfunction in HSRP that caused a subset of the active HSRP interfaces to stop passing traffic even though the interface state did not change.  The root cause of the malfunction has not been identified

Discussion

How much failed?

The HSRP malfunction affected a small subset of HSRP enabled interfaces on one of two High Availability board routers located in Uber’s Sydney data centre.  The issue only impacted a small number of customers, and the level of impact depended on the services that the customer was receiving, the location of the internet based users attempting to access their sites, the architecture of their solutions, and whether they were receiving services via an interface impacted by the issue.

What services were affected?

The HSRP failure caused different issues for customers, depending on the services the customers were receiving:

• The vast majority of customers saw no impact.  The services they were receiving were not provided by interfaces affected by the HSRP failure.
• For some customers the impact was a minor degradation in service whereby a small number of internet users would not have been able to make use of the customer’s hosted services.
• For a very small number of customers the HSRP failure caused a total loss of services until the malfunction was rectified.

While Uber runs in border routers in High Availability mode using HSRP physical connections to upstream carriers can only be terminated on a single interface simply because they are presented as a single piece of physical media.  The Sydney border router that experienced the malfunction was terminating transit connections from Extel and Vocus.  Services to internet users who’s traffic passed through either Extel or Vocus via Uber’s Sydney data centre would have been disrupted by the malfunction.  Users accessing Uber’s Sydney data centre via AAPT, Internode or Pipe Networks would have not experienced any degradation.

Customers with services directly connected to the border router interfaces that experienced the HSRP malfunction would have experienced a total loss of connectivity for the duration of the malfunction.

Should Monitoring have pointed to the Real Problem Immediately?

Uber monitors the performance of router services directly.  This monitoring can normally detect router or network related problems almost instantaneously by examining the router syslog output.  In the case of the HSRP malfunction the interface affected did not change state, and therefore did not generate syslog that may have pointed to an issue.  Additionally the malfunction only affected a small subset of the total number of HSRP enabled interfaces, and the malfunction only occurred on a single router.  It is not clear that any level of monitoring that could be enabled by Uber would have directly detected the issue.

Corrective Actions

In order to support routine growth Uber has already commenced a project to replace the existing border routers in Canberra and Sydney. This work is scheduled for completion in June.  A replacement of the border routers will resolve this issue.

Do you run Microsoft software in your business?

Then it makes sense to use a Windows Web Server, made by Microsoft, for Microsoft.

Would you like to easily configure your web and email services?

Having your own server is great, but configuring web and email services can be a headache. Get the benefits of a VPS and easy web deployment through Plesk, which provides an easy to use web-based control panel for web services. You can configure domains, create email accounts, manage FTP and web settings - the list goes on.

Do you want to install a blog, Content Management System, shopping cart for eCommerce, forum, gallery or wiki?

Windows Web Server lets you create your own custom server environment, and makes it easy to combine powerful Windows software such as ASP .Net with popular open source solutions like PHP and MySQL. This allows you to run popular web applications such as WordPress, Joomla, ZenCart, phpBB, Drupal and much more!

Under the Windows OS you also get true multi-threading so your applications run faster. Plus, you can do all this through the familiar Windows GUI – no need to struggle with a command line interface and Unix commands.

Do you want to save your business time and money, without compromising on performance?

• Host multiple websites on one Windows Web Server, using Plesk to manage your Internet Information Service.
• Windows Web Servers run on Microsoft Hyper-V. This gives you all the benefits of a dedicated server at a fraction of the cost.
• Windows Web Servers are highly scalable, allowing you to save money by starting small and growing your solution at the same pace as your business.
• Windows Web Servers can be provisioned in real time, meaning you can purchase your server and have your website or application up and running within a couple of hours.
• Windows Web Servers are hosted on enterprise grade hardware in our Australian data centre.
• Windows Web Servers come with industry leading Windows Server 2008 R2 Standard.

Check out our introductory offer!

For a limited time only we are offering you the chance to sign up for a Windows Web Server with the first month completely free!

Click here or call 13 UBER (13 8237) to save up to $427!!!!

Outbound email from Uber Shared Hosting customers using mail clients (ie Outlook etc, not Webmail) normally works as follows:

• The User creates an email on their personal computer, mobile device etc using their mail client (Outlook, Mail etc).
• The email leaves the user’s computer/device and travels to the Shared Hosting Server account provided by Uber (both Plesk and cPanel shared hosting servers) via smtp.
• The Shared Hosting Server sends this email to the recipient via Uber’s internet connections.  Uber filters outbound emails as they leave our network for spam and viruses to ensure that the Uber network is not a source of malicious traffic to the internet.

The one-to-one relationship between the User and the Shared Hosting Server account is authenticated by the email address and mail password.

There are many ways for spammers to exploit this system.  One common method is to obtain or guess the user’s email addresses and then attempt to guess the password.  The user’s email address can be obtained from many different places on the internet, including straight off the user’s website itself.  An email address can be guessed because shared web accounts typically set up common email addresses such as admin@ or support@.  Weak passwords can be guessed using brute force dictionary attacks.

Web servers have features built in to prevent such attacks:

• Brute force attacks are sometime defeated by locking accounts if too many failed logins occur from the same IP address.  Spammers can overcome this approach by using a large number of coordinated servers to ensure that each guess comes from a different IP address.
• Testing passwords for strength by checking for non-standard characters or running the password against a dictionary before accepting it.  These measures are often effective but unfortunately they only been implemented seriously by a number of Shared Hosting control panel developers in the last few years.  Accounts that have been in existence for many years are likely to have been established before rigorous password testing was implemented by control panel vendors.

Finally when people hear that passwords are tested against a ‘dictionary’, they may think ‘Oxford’, ‘Webster’ or ‘Macquarie’.  The password dictionary we used to conduct our audit for weak passwords has over fourteen million entries.  It includes variations on ‘password’ such as ‘P@5sw0rd!’ and many non-words such as ‘1234567’.  Unless your password contains a mixture of upper and lower case characters, special characters and numbers, then it is likely to not be very strong.

The Incident

Commencing on Monday 12 March, Engineers noticed an increase in spam being generated from both Plesk and cPanel Shared Hosting Servers. Due to the nature of low cost hosting there are often minor spam outbreaks generated from within the Uber shared hosting fleet.  Uber employ outbound mail filtering to catch spam before it leaves our network and alert our Engineers who then rectify the source accounts. However the increase on Monday was significantly higher than normal and began to overload our outbound mail filters.

Initial investigations into the spam source revealed a large number of spamming email accounts. Engineers investigated and discovered that the spamming accounts were located on both cPanel and Plesk Shared Servers.  It appeared that the spammers had obtained copies of both the email address and passwords of the email accounts.  As the spamming accounts were identified the passwords were changed, thereby stopping the spam, and the customers notified.

Over the week more and more accounts began spamming causing the number of spam emails to increase dramatically to the point that the Mail Transfer Agents employed to forward email to the internet were at times overwhelmed with the volume of mail, causing delays in sending and receiving mail.  Later in the week Uber began to discover scripts loaded on a small number of accounts that were also starting to generate large volumes of email traffic.  As each account is discovered the password is reset (and if relevant the script removed) so the account can no longer be used to send spam.

Investigations have shown that the only common factor in all these events is the use of weak account passwords.

Analysis of logs shows:

• No obvious password brute-forcing in progress presently which indicates the reconnaissance phase is probably over and now the spammers are in exploit phase.
• A very large number of servers initiating spamming smtp connections, suggesting that a large netbot is probably being used to target the Uber infrastructure.

Unfortunately these two facts means there are no obvious high-threat nodes that Uber can block to reduce the volume of the event.

Analysis of historical logs shows no obvious high volume brute-force IP addresses which reinforces our view that the source of the traffic is a large netbot, indicating a sophisticated spammer.

Root Cause

The root cause of the delays in email delivery over the last week is the use by spammers of Share Hosting email and other accounts with weak passwords to send large volumes of outbound spam. This spam clogs the mail filters, slowing the email at points during the day.

Risk to Customers

The passwords that the spammers are guessing are mail passwords, not account passwords.  There is no evidence of spammers gaining access to customer’s billing details.  However with access to the email address/password combination it possible to read a user’s email.  Based on our analysis of the logs we do not see any evidence of spammers accessing customers email, but this is always a possibility.  However this event highlights the importance for internet users to ensure they use strong passwords in their dealings on the web.

Short Term Fix

A number of methods to aid with rapid detection and email delivery management have now been employed to minimise the impact to customers allowing Uber to very quickly identify and remediate accounts that are in use by spammers.

Medium Term Fix

Planning is underway to perform a large-scale password update for customers identified as having weak passwords.  We will be shortly contacting these customers.  If you believe that you have a weak password for any of your accounts presently held with AussieHQ or Jumba it is suggested you change it immediately.

Long Term Fix

There are at least two possible long term fixes:

• Investigate methods for retrofitting password strength testing tools to old control panel distributions (upgrading the control panel software will not retrospectively force the passwords to be re-tested).
• Implement routine password strength audits to identify Users who have entered weak passwords and contact them to upgrade their strength.

Other options continued to be investigated.