Onboarding One of Norway’s Largest ISPs with Faster, More Stable Infrastructure
When you upgrade your IT infrastructure, enhanced performance and new features are typically top of mind. At the same time, stability and reliability are crucial. Nobody wants to move to new equipment only to discover hardware or software issues resulting in system failures and downtime. Refreshing your infrastructure is challenging at the best of times, but it’s especially tricky following a merger—or any time when companies consolidate systems and processes. To succeed, you must find the right combination of cutting-edge and tried-and-tested technology.
I’m the Operations Team Leader at Bitpro, a Norwegian provider of cloud-based IT and security solutions and access-independent wired and wireless broadband since 2007. Our 90 employees in Oslo, Bergen, and Kristiansand support and deliver services to over 3,700 business customers, ranging from SMEs to larger companies.
Over the years, Bitpro expanded our offerings through a series of mergers and acquisitions. We strengthened our investment in communication solutions for larger companies with high requirements for delivery performance and flexibility. These moves increased our competitiveness and product portfolio. But while our product line got more comprehensive, our IT infrastructure got more complicated, requiring us to rethink our internal organization.
Standardizing IT Operations
Bitpro has a big delivery and operations department that is split between customers and internal services. As the team leader for IT operations, I handle the day-to-day details for our internal operations, while other team members are dedicated to customer services. I have a long history leading IT operations at our sister company, NextGenTel, and Bitpro wanted me to continue this work here.
Shortly after I came to Bitpro, NextGenTel decided to migrate its IT infrastructure from their previous operations partner to Bitpro. That was a big deal, because NextGenTel is one of Norway's biggest ISPs. But onboarding NextGenTel meant increased workloads for our infrastructure.
We had to figure out how many servers we needed to deploy for a successful migration based on improved technology and forecasted demand. Space, heat, and power consumption were important to us, and because we are spread across locations, ease of maintenance was also a consideration. We also needed to choose a vendor that could offer the best balance of price, performance, and support.
Learning from a Previous Refresh
Several years ago, we underwent an upgrade to standardize our infrastructure, which supports Hyper-V and Oracle for our ISP business. For compute, we chose six HPE ProLiant DL325 servers. The new servers proved to be reliable and stable. We liked that the servers used AMD processors, which allowed us to get the same performance with fewer servers, and they also offer more cores per server than competitive machines at the same price. We also replaced our storage with three HPE Nimble Storage arrays, and we were happy with the level of service and support from HPE around monitoring.
The NextGenTel migration was a massive project. We needed to move 250 different types of virtual servers, including database servers and nearly 80 terabytes of data, to our infrastructure. Since we had a good experience with HPE, it made sense to turn to them for a solution that would increase the performance and availability of business-critical applications while reducing complexity and total cost of ownership.
Migrating NextGenTel’s Infrastructure to Our Data Center
We had already augmented our SAN in preparation for the migration and had initially planned to install 10–15 servers. Our initial estimate didn't factor in how much performance we could get from our AMD-based machines. We completed many performance tests around the initial implementation. When we showed NextGenTel the results—that moving to the HPE ProLiant DL325 servers would maintain a high level of performance—it proved to be a big factor in NextGenTel selecting Bitpro as an operations partner.
We ended up installing eight new HPE ProLiant DL325 servers that co-exist with the original six servers on our premises and connecting them to our existing SANs. We also installed VMware on these servers because NextGenTel mostly runs Linux, whereas Bitpro is primarily a Windows shop. We then added a fourth HPE Alletra to replace the oldest HPE Nimble Storage array.
Having migrated the servers, we connected them to our SANs. Most of the NextGenTel servers are running the same OS, which is quite good for the deduplication of the SAN—we got a high deduplication rate on each partition. This result meant we could fit more data on it than anticipated and had more space afterward, potentially delaying the need to improve or extend the SAN later. If the deduplication rate had been lower, we would probably need to extend the SAN earlier than anticipated.
We use mechanical hard drives and SSDs at our data center, but NextGenTel also had many Oracle databases running with high input/outputs. HPE Alletra handles those SSDs, so we could get those high IOPS as well.
We had previously used IBM and HPE 3PAR SANs, but HPE Nimble and HPE Alletra are far easier to manage using HPE InfoSight, an AI-powered portal. It used to be challenging to troubleshoot customers’ performance issues using traditional SANs and network management software, but HPE InfoSight and HPE Nimble make it easier for us to find and eliminate bottlenecks. HPE InfoSight also allows us to troubleshoot our storage arrays remotely, which is especially important to Bitpro as our data center is off-site. Combined with remote server management from HPE Integrated Lights Out (iLO), we can manage all our infrastructure, including our third-party backup data center, without leaving our offices. That’s invaluable in a distributed environment with multiple locations such as ours.
We also leverage HPE Smart Update Manager, an intuitive graphical interface that guides users through installing HPE ProLiant Support Packs. Our servers run 24/7, and we can’t risk automatic updates that might compromise our systems. HPE Service Pack bundles all the drivers we need. It performs dependency and validity checks, ensuring that everything works together so we don’t have to test and install every driver individually.
Stability, Reliability, and Capacity for Future Growth
Our hybrid model offers a lot of space and data functionality, and the data deduplication rate is better than anticipated, which allows us more space for more data. NextGenTel is running lower server loads and using fewer processor cores on AMD architecture than on its older machines. They’re also using less storage, allowing us to reduce our energy consumption and operate more sustainability. HPE Nimble Storage and HPE Alletra’s advanced deduplication algorithms eliminate the resulting redundant data, making every partition smaller.
By combining HPE ProLiant DL325 servers and HPE Nimble/Alletra storage arrays, we're able to optimize performance for any workload. This has given Bitpro the tools to successfully navigate any expansions, whether through mergers or new customer loads. We gained the perfect platform for the NextGenTel migration and now have a combined data center that serves Bitpro’s thousands of business and enterprise customers while providing the infrastructure for NextGenTel’s consumer broadband and IPTV products. Together, our two companies have led to a simpler, more efficient, and more cost-effective infrastructure.