HYCU, Inc. presenting at #CFD17 from Tech Field Day mentioned that "Everyone knows they have to move to cloud." True enough. We're all there or going there. Maybe not for everything, but for lots of things. Much of that cloud consumption is via SaaS. This hit me in a personal way. My company is entirely in the cloud. We use many SaaS services, and more every week. How are we backing up the data we've entrusted to these various SaaS providers? This is the challenge HYCU, already an established data protection vendor, has taken on with their R-Cloud product. R-Cloud is addressing 3 specific challenges that SaaS consumers face. Operational data loss (someone screwed up or something broke), cyber events & outages (someone bad did something or something broke), and data compliance & sovereignty (regulatory issues). R-Cloud is doing this in a low-code way, so that ANY SaaS service can map their data to HYCU's R-Cloud. The data is then available to be managed like you'd leverage any backup tool you've been using for years--retention policies, recovery, etc. From a SaaS consumer perspective, there's an R-Cloud ecosystem that's expected to grow. I could use that to see if my SaaS tool is in there. I might even use it as metadata for a SaaS buying decision. Is this service I'm considering backed up by HYCU? A point in the favor of that service. No? Maybe a negative or neutral point. Another question is whether as a SaaS consumer I'd want to use R-Cloud to build backup for a SaaS, rather than wait for my SaaS provider to build a backup integration for me. Even if the backup platform is low-code, i.e. "easy to use", a SaaS provider is going to be far more interested in adding sexy, visible new features to their tool. Backup is neither sexy nor visible unless there's actual data loss. But what investor wants to hear from the SaaS service they invested in that they've got data protection now? That puts the ball back in my court, perhaps--would I (could I?) use HYCU's R-Cloud to backup an important-to-my-business dataset living in my chosen SaaS? I might. I just might. Would you? If you could use a low-code tool to create a robust backup scheme for any SaaS product you consume, is that interesting? Or would you keep doing what most of us have been doing--praying to the cloud gods that there's never a catastrophic outage that kills our data or an employee with lots of access that gets hacked (or simply has a bad day)? I'm thinking a lot more about SaaS data protection all of a sudden. The cloud gods are capricious. #business #data #tech #cloud #compliance #security #dataprotection
Ethan Banks’ Post
More Relevant Posts
-
"You don't have to keep the lights on if the lights are automated." - Karl Newell @ Network Automation Forum #AutoCon0. What's Karl getting at? The problem engineers are having getting the time & support they need from their businesses to invest in network automation. If engineers are constantly busy performing network maintenance & changes to support the business, they don't have time to automate the network. But if they did have that time, the business would gain a strategic advantage. Faster time to market. Reduction in human error. Simpler recovery from changes that didn't go as planned. I view this as not primarily an engineering/IT problem--it's a business problem.
-
"The source of truth can't be the network itself." - Nick Thompson I 100% agree with Nick who shared via a panel discussion at Network Automation Forum #AutoCon0 today. I didn't agree with this POV once upon a time, because what could be more "true" than the network itself? Then it dawned on me that a source of truth represents what things *should be* as opposed to *what they actually are*. When I got that into my head, how I thought about network management shifted.
-
If you're at Network Automation Forum #AutoCon0, Andrew Conry-Murray and I are in the Larkspur room. Come by and say hi! And if you're willing, we'd like to put you on camera for a compilation video we're making. We're asking everyone the same set of questions... - What do engineers get wrong on their first network automation attempt? - Does your company dedicate money or time to support network automation? - What’s holding back the industry from progressing on network automation? - Should the long-term goal for network automation be self-service? - If you could tell network equipment vendors one thing related to automation, what would it be?
-
Software defined storage provider Qumulo offers a file system for unstructured data you can run where you like--on premises using commodity hardware or in the cloud. Qumulo emphasizes massive scalability. - Qumulo customers average ~1PB of data. - 90% of operations talk 1 millisecond or less. - Qumulo tends to sit on a hybrid platform, a mix of SSD and HDD. Or even cached to RAM, using a predictive algorithm to populate the cache. - Key industries for Qumulo include media & entertainment, healthcare, fintech, manufacturing, life sciences, and energy. Qumulo was new to me until last week when Ryan Farris briefed Packet Pushers. Qumulo has announced the latest version of their Scale Anywhere platform. Qumulo is unifying their products to make it easier for customers to manage their massive datasets, no matter where those datasets reside. There are several components to this unification-focused announcement. 1. In April 2023, Qumulo launched Azure Native Qumulo (ANQ). The big idea is to present the Qumulo file system on the front side to hosts, while using Azure blob storage on the backend. If you have a client that wants to talk Qumulo, but you don't want to pay what you'd normally have to pay to get a proper file system in the cloud (such as Netapp on AWS), ANQ is a good answer. You get a cloud-based file system using cheap Azure blob storage and save money on your cloud bill. Qumulo claims, "ANQ is almost 80% less expensive than the closest alternative, and is comparable to the fully burdened on-premises cost of file storage." ANQ is a key component of Qumulo Scale Anywhere. 2. Qumulo Global Namespace has been announced, offering a unified file system no matter how geographically distributed your files are. Got files on the edge, on premises, as well as in the cloud? No worries. Treat them as one massive file system. It's a CDN model. Let's say SFO is a hub, and we're sharing files across Global Namespace to spokes SEA & NYC. File catalogs are shared to SEA & NYC such that you can browse them. When you select a file, the file needed is copied to the spoke. From there, it's locally cached. You only deal with the performance penalty during initial retrieval. Global Namespace is in private preview until January 2024, after which it will move to Public Preview. 3. Qumulo Nexus offers a single pane of glass of your Qumulo instances. Nexus shows you the things you'd expect to see to help you monitor utilization of your file systems. At a glance, it's perfectly adequate, but not breaking any new ground. You've used tools like it before. Graphs and stats and dashboards, oh my. 4. Qumulo One is how you track your Qumulo spend. You get a unified licensing interface in One, and Qumulo claims this is simple to manage. Qumulo Scale Anywhere: https://lnkd.in/gjjJYQ9r ANQ pricing: https://lnkd.in/gYrGcK3m #storage #cloudstorage #petabyte
-
-
Recording this Day Two Cloud episode with Frederic Van Haren was highly educational for me. If you're any sort of infrastructure engineer or architect, you want to listen to this one. Frederic understands the processes of ML & AI and the computing stacks required to make them go. A lot of words around AI & ML get glossed over regarding modeling, metadata, training, and so on in general technical conversations. Frederic knows what all these terms mean, the impact each of these has on AI & ML, and how you as an infrastructure pro have to think about IT resources as a result. #podcast #ai #ml #hpc #infrastructure #infrastructureengineer
Have you ever wondered what sits behind all this #AI craziness? What powers #ChatGPT or #DALLE? How would you go about setting up a private AI instance? Ethan Banks and I were fortunate to have Frederic Van Haren on the most recent episode of Day Two Cloud to discuss these questions and much more. Frederic has been doing AI and HPC for over a decade and he brings a wealth of experience to the episode. Def worth a listen if you're curious about the nuts and bolts behind AI. https://lnkd.in/eZusmv-V
-
-
If you're not subbed to Markku's RSS, worth a look! Solid informational & educational posts for network engineers. I've been subbed for years. Markku's blog: https://majornetwork.net/
This year I have written a few posts about #DHCP, the Dynamic Host Configuration Protocol. Why not add one more? Here is a scenario where it looks like the DHCP servers are slow to respond, but the actual case is a bit more convoluted: https://lnkd.in/ejAUr4Sz
Technical Educator and Content Creator | Microsoft MVP 6x | HashiCorp Ambassador 4x
5mo"Backup is neither sexy nor visible unless there's actual data loss." True dat. I love the idea of a vendor agnostic backup platform that provides simple integration tools. One thing we didn't look at was cost. I'm very curious what licensing model is applied to their solution. Probably data protected by GB?